Microsoft Programmer's Library 1.3

home *** CD-ROM | disk | FTP | other *** search

/ Microsoft Programmer's Library 1.3 / Microsoft_Programmers_Library.7z / MPL / net / sqllang.txt < prev next >

Wrap

Text File | 2013-11-08 | 674.8 KB | 21,330 lines

Microsoft SQL Server - Language Reference ──────────────────────────────────────────────────────────────────────────── Microsoft (R) SQL Server - Language Reference The SYBASE (R) SQL Server database for PC networks VERSION 1.1 ──────────────────────────────────────────────────────────────────────────── for the MS (R) OS/2 Operating System Microsoft Corporation Information in this document is subject to change without notice and does not represent a commitment on the part of Microsoft Corporation. The software described in this document is furnished under a license agreement or nondisclosure agreement. The software may be used or copied only in accordance with the terms of the agreement. It is against the law to copy the software on any medium except as specifically allowed in the license or nondisclosure agreement. No part of this manual may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, for any purpose without the express written permission of Microsoft Corporation. (C) 1990 Microsoft Corporation and SYBASE, Inc. All rights reserved. Printed in the USA. 1 = Microsoft, MS, MS-DOS, and the Microsoft logo are registered trademarks of Microsoft Corporation. IBM is a registered trademark of International Business Machines Corporation. SYBASE is a registered trademark of SYBASE, Inc. TRANSACT-SQL and DB-LIBRARY are trademarks of SYBASE, Inc. Document Number: SY10231-0290 OEM-D/0788-1Z Table of Contents ──────────────────────────────────────────────────────────────────────────── Before You Begin Manual Overview How to Use This Guide Notational Conventions Finding Further Information Chapter 1 TRANSACT-SQL Statements Aggregate Functions ALTER DATABASE ALTER TABLE Batch Queries BEGIN...END BEGIN TRANSACTION BREAK Browse Mode CHECKPOINT Comments COMMIT TRANSACTION COMPUTE Clause CONTINUE Control-of-Flow Language Conversion Function CREATE DATABASE CREATE DEFAULT CREATE INDEX CREATE PROCEDURE CREATE RULE CREATE TABLE CREATE TRIGGER CREATE VIEW Datatypes Date Functions DBCC DECLARE DELETE DISK INIT DISK REFIT DISK REINIT DROP DATABASE DROP DEFAULT DROP INDEX DROP PROCEDURE DROP RULE DROP TABLE DROP TRIGGER DROP VIEW DUMP DATABASE DUMP TRANSACTION EXECUTE Expressions Functions GOTO GRANT GROUP BY and HAVING Clauses Identifiers IF...ELSE INSERT Joins KILL LOAD DATABASE LOAD TRANSACTION Mathematical Functions Null Values ORDER BY Clause Parameters PRINT RAISERROR READTEXT RECONFIGURE RETURN REVOKE ROLLBACK TRANSACTION Row Aggregate Functions SAVE TRANSACTION Search Conditions SELECT SET SETUSER SHUTDOWN String Functions Subqueries System Functions System Procedures Text/Image Datatypes Text/Image Functions TRUNCATE TABLE UPDATE UPDATE STATISTICS USE Variables (Local and Global) Views WAITFOR WHERE Clause WHILE Wildcard Characters WRITETEXT Chapter 2 System Procedures sp_addalias sp_addgroup sp_addlogin sp_addtype sp_addumpdevice sp_adduser sp_bindefault sp_bindrule sp_changedbowner sp_changegroup sp_commonkey sp_configure sp_dboption sp_defaultdb sp_depends sp_diskdefault sp_dropalias sp_dropdevice sp_dropgroup sp_dropkey sp_droplogin sp_droptype sp_dropuser sp_foreignkey sp_help sp_helpdb sp_helpdevice sp_helpgroup sp_helpindex sp_helpjoins sp_helpkey sp_helprotect sp_helpsql sp_helptext sp_helpuser sp_lock sp_logdevice sp_monitor sp_password sp_primarykey sp_rename sp_renamedb sp_spaceused sp_unbindefault sp_unbindrule sp_who Chapter 3 Utility Programs bcp bldmastr console defncopy isql sqlservr Appendix A System Tables sysalternates (all databases) syscolumns (all databases) syscomments (all databases) sysconfigures (master database only) syscurconfigs (master database only) sysdatabases (master database only) sysdepends (all databases) sysdevices (master database only) sysindexes (all databases) syskeys (all databases) syslocks (master database only) syslogins (master database only) syslogs (all databases) sysmessages (master database only) sysobjects (all databases) sysprocedures (all databases) sysprocesses (master database only) sysprotects (all databases) syssegments (all databases) systypes (all databases) sysusages (master database only) sysusers (all databases) Appendix B The pubs Sample Database B.1.1 Rules Trigger Stored Procedure View Defaults Appendix C TRANSACT-SQL Keywords Index Before You Begin ──────────────────────────────────────────────────────────────────────────── Manual Overview This reference manual describes each statement and command of TRANSACT-SQL(tm), an enhanced version of the SQL relational database language. It also describes SQL Server system procedures and utility programs. Before using this manual, you should have read SQL Server Administrator's Guide for information on using the SQL Server Administration Facility (SAF). You should also be familiar with TRANSACT-SQL or another version of SQL. To learn TRANSACT-SQL, see SQL Server Learning TRANSACT-SQL. How to Use This Guide Topics in each chapter are listed in alphabetical order. The topics include each statement, or command, with complete syntax as well as general topics. For example, to find the complete syntax of the SELECT statement, look under "SELECT" in Chapter 1. To read about identifiers, look under "Identifiers" in Chapter 1. This manual covers the following topics: ──────────────────────────────────────────────────────────────────────────── Chapter 1 TRANSACT-SQL, SQL Server's enhanced version of the SQL database language Chapter 2 SQL Server system procedures, which are procedures supplied with SQL Server and executed from the SQL Server Administration Facility (SAF) or isql Chapter 3 SQL Server utility programs, which are supplied with SQL Server and executed directly from the operating system Appendix A SQL Server system tables, which are supplied with SQL Server or created when the CREATE DATABASE statement is executed Appendix B The sample database, pubs (used in examples throughout the manual) Appendix C TRANSACT-SQL keywords ──────────────────────────────────────────────────────────────────────────── The SQL Server Language Reference is part of the SQL Server documentation set. Several manuals in this set are fully illustrated with examples from the online sample database, pubs. (See Appendix B, "The pubs Sample Database.") Ask your System Administrator how to get a clean copy of the pubs database for use in learning TRANSACT-SQL and the system procedures. Notational Conventions Throughout this manual, the following conventions are used to distinguish elements of text: ╓┌─────────────────────────────────┌─────────────────────────────────────────╖ Convention Purpose ──────────────────────────────────────────────────────────────────────────── UPPERCASE Represents statement and clause names, functions, macros, and any other portions of syntax that must appear exactly as shown. MIXEDcase Shows abbreviations for keywords in syntax. The uppercase letters show the required portion of the keyword; the lowercase letters show the optional Convention Purpose ──────────────────────────────────────────────────────────────────────────── lowercase letters show the optional portion. SMALL CAPS Represent key names such as CTRL. bold Represents stored procedures, system procedures, triggers, defaults, rules, utility programs, and commands. italic Represents database names, table names, view names, column names, datatypes, index names, pathnames, filenames, and variables that appear in text. monospace Represents examples, screen output, program code, and error messages. [brackets] Enclose optional items. Type only the information within the brackets, not the Convention Purpose ──────────────────────────────────────────────────────────────────────────── information within the brackets, not the brackets themselves. {braces} Enclose required items. Type only the information within the braces, not the braces themselves. | (vertical bar) Separates items inside a set of braces or brackets. The vertical bar means you must choose one and only one item. ... (ellipsis) Means that you can repeat the previous item as many times as you like. <execute> Executes one or more SQL statements. (In the SQL Server Administration facility, SQL statements are executed by pressing the CONTROL+E keys. In the isql program, SQL statements are executed with the go Convention Purpose ──────────────────────────────────────────────────────────────────────────── SQL statements are executed with the go command.) Finding Further Information The following manuals describe SQL Server and are included as part of the standard documentation set: ──────────────────────────────────────────────────────────────────────────── SQL Server Installation Guide A guide to installing and setting up SQL Server SQL Server Learning TRANSACT-SQL A guide to learning and using TRANSACT-SQL SQL Server System Administrator's Guide A guide for SQL Server System Administrators and for using the SQL Server Administration Facility for database queries SQL Server Programmer's Reference A reference to DB-LIBRARY(tm), which is a set of C routines and macros that allow your application to interact with SQL Server SQL Server Quick Reference A quick reference guide to TRANSACT-SQL ──────────────────────────────────────────────────────────────────────────── Chapter 1 TRANSACT-SQL Statements ──────────────────────────────────────────────────────────────────────────── This chapter consists of alphabetized reference pages, one for each TRANSACT-SQL statement plus special topics such as datatypes, identifiers, and functions. Some SQL statements that are particularly complex (such as SELECT) are broken down into individual subsections. For example, there are separate reference pages on the COMPUTE, GROUP BY, and HAVING clauses of the SELECT statement. The following statements, commands, and topics are covered in this chapter: ──────────────────────────────────────────────────────────────────────────── Aggregate Functions Return summary values. ALTER DATABASE Increases the amount of disk space allocated to a database. ALTER TABLE Adds new columns to an existing table. Batch Queries A set of SQL statements submitted together and executed as a group, one after the other. BEGIN...END Encloses a series of SQL statements so that control-of-flow language, such as IF...ELSE, affects the performance of the whole group. BEGIN TRANSACTION Marks the starting point of a user-specified transaction. BREAK Controls operation of statements in a WHILE loop. Browse Mode Supports the ability to perform updates while viewing data. CHECKPOINT Forces all dirty pages in the current database to be written to the disk. Comments Provide information about SQL statements, statement blocks, and stored procedures. COMMIT TRANSACTION Marks the ending point of a user-defined transaction. COMPUTE Clause Generates summary values in a SELECT statement with row aggregate functions. CONTINUE Controls operation of statements in a WHILE loop. Control-of-Flow Language Controls the flow of execution of SQL statements, statement blocks, and stored procedures. Conversion Function Converts expressions of one datatype to another datatype. CREATE DATABASE Creates a new database. CREATE DEFAULT Specifies a value that will be inserted in a column if no value is explicitly supplied at insert time. CREATE INDEX Creates indexes. CREATE PROCEDURE Creates a stored procedure that can take one or more user-supplied parameters. CREATE RULE Specifies the domain of acceptable values for a particular column or for any column of a specified user datatype. CREATE TABLE Creates new tables. CREATE TRIGGER Creates a trigger, which is a special kind of stored procedure often used for enforcing integrity constraints. CREATE VIEW Creates views. Datatypes Specify data characteristics of columns, stored procedure parameters, and local variables. Date Functions Manipulate datetime values. DBCC Checks the logical and physical consistency of a database. DECLARE Declares the name and type of local variables for a batch or procedure. DELETE Removes rows from a table. DISK INIT Reserves and formats physical storage for database devices. DISK REFIT Restores a damaged master database. DISK REINIT Part of the procedure that restores a damaged master database. DROP DATABASE Removes one or more databases from SQL Server. DROP DEFAULT Removes a user-specified default. DROP INDEX Removes an index from the database. DROP PROCEDURE Removes user-created stored procedures from the current database. DROP RULE Removes a user-specified rule. DROP TABLE Removes a table definition and all data, indexes, triggers, and permission specifications for a table from the database. DROP TRIGGER Removes a trigger. DROP VIEW Removes views from the database. DUMP DATABASE Makes a backup copy of the database and the transaction log. DUMP TRANSACTION Removes the inactive part of the transaction log and makes a backup copy of it. EXECUTE Runs a system procedure or a user-defined stored procedure. Expressions Used as variables and constants in many SQL statements. Functions Return special information from the database. GOTO Causes an unconditional branching to a user-defined label. GRANT Assigns permissions to users. GROUP BY and HAVING Clauses Divide a table into groups. Identifiers Name database objects─databases, tables, views, columns, indexes, triggers, procedures, defaults, rules, and so on. IF...ELSE Impose conditions on the execution of an SQL statement. INSERT Adds new rows to a table or view. Joins Compare two or more tables (or views). KILL Kills a process. LOAD DATABASE Loads a backup copy of a user database and its transaction log that was created with DUMP DATABASE. LOAD TRANSACTION Loads a backup copy of the transaction log. Mathematical Functions Return values commonly needed for operations on mathematical data. Null Values Mark columns having an unknown value. ORDER BY Clause Returns query results in the specified column(s) in sorted order. Parameters Values supplied to a stored procedure. PRINT Prints a user-defined message on the user's screen. RAISERROR Prints a user-defined error message on the user's screen and sets a system flag to record that an error has occurred. READTEXT Reads text and image values, starting from a specified offset and reading a specified number of bytes. RECONFIGURE Part of the procedure that sets configuration options, which control various aspects of SQL Server's memory allocation and performance. RETURN Exits unconditionally from a query or procedure. REVOKE Revokes permissions from users. ROLLBACK TRANSACTION Rolls back a user-specified transaction to the last savepoint inside a transaction or to the beginning of a transaction. Row Aggregate Functions Generate summary values that appear as additional rows in the query results. SAVE TRANSACTION Sets a savepoint within a transaction. Search Conditions Set the conditions in a WHERE or HAVING clause. SELECT Retrieves rows from the database. SET Sets SQL Server query-processing options for the duration of the user's work session, or inside a trigger or stored procedure. SETUSER Impersonates another user. SHUTDOWN Brings the system to a halt. String Functions Perform various operations on binary data, character strings, or expressions, including concatenation. Subqueries Nest a SELECT statement inside a SELECT, INSERT, UPDATE, or DELETE statement, another subquery, or anywhere an expression is allowed (if it returns a single value). System Functions Return special information from the database. System Procedures Stored procedures that SQL Server supplies to update and report from the system tables. Text/Image Datatypes The text datatype stores extremely long (up to 231) byte strings of printable characters. The image datatype stores extremely long (up to 231) bytes of hexadecimal-encoded binary data. Text/Image Functions Return values commonly needed for operations on text and image data. TRUNCATE TABLE Removes all rows in a table as quickly as possible. UPDATE Changes data in existing rows, either by adding new data or modifying existing data. UPDATE STATISTICS Updates information about the distribution of key values in specified indexes. USE Changes the current database. Variables (Local and Global) Defined entities that are assigned values. Views Provide an alternate way of looking at data in one or more tables. WAITFOR Specifies a specific time, a time interval, or an event for the execution of a statement block, stored procedure, or transaction. WHERE Clause Sets the conditions in a WHERE clause. WHILE Sets a condition for the repeated execution of a statement or statement block. Wildcard Characters Used with the LIKE keyword to represent any character in a string when searching for a char, varchar, or datetime value. WRITETEXT Permits nonlogged, interactive updating of an existing text field. ──────────────────────────────────────────────────────────────────────────── Aggregate Functions ──────────────────────────────────────────────────────────────────────────── Function Returns summary values. Aggregate functions are SUM, AVG, COUNT, COUNT(*), MAX, and MIN. They can be used in the select list or the HAVING clause of a SELECT statement or subquery, and often appear in a statement that includes a GROUP BY clause. (A similar type of aggregate function, called a row aggregate function, is used in the COMPUTE clause.) Syntax Aggregate functions have the following syntax: aggregate_function ([DISTINCT] expression) Aggregate functions, their individual syntax, and the results they produce are shown in the following table. Note that expression is usually a column name. DISTINCT is optional for SUM, AVG, and COUNT but must not be used with COUNT(*). COUNT(*) takes no parameters. The numeric datatypes are int, smallint, tinyint, float, and money. ╓┌───────────────────┌────────────────────────┌──────────────────────────────╖ Aggregate Function Parameters Result ──────────────────────────────────────────────────────────────────────────── AVG ([DISTINCT] expression) The average of the [distinct] values in the numeric column COUNT ([DISTINCT] expression) The number of [distinct] non-null values in the column COUNT (*) The number of selected rows MAX (expression) The highest value in the expression MIN (expression) The lowest value in the expression Aggregate Function Parameters Result ──────────────────────────────────────────────────────────────────────────── AVG ([DISTINCT] expression) The average of the [distinct] SUM ([DISTINCT]expression) The total of the [distinct] values in the numeric column ──────────────────────────────────────────────────────────────────────────── Examples A. select avg(advance), sum(ytd_sales) from titles where type = "business" Example A calculates the average advance and the sum of year-to-date sales for all business books. Each of these aggregate functions produces a single summary value for all of the retrieved rows. B. select type, avg(advance), sum(ytd_sales) from titles group by type When used with a GROUP BY clause, aggregate functions produce single values for each group, rather than for the whole table. The statement in example B produces summary values for each type of book. C. select count(distinct city) from authors Example C finds the number of different cities in which authors live. D. select type from titles group by type having count(*) > 1 Example D lists the types in the titles table but eliminates the types that include only one book or none. E. select pub_id, sum(advance), avg(price) from titles group by pub_id having sum(advance) > $25000 and avg(price) > $15 Example E groups the titles table by publisher, and includes only those groups of publishers who have paid more than $25,000 in total advances and whose books average more than $15 in price. Options DISTINCT Eliminates duplicate values before an aggregate function is applied. DISTINCT is optional with SUM, AVG, and COUNT. column_name The name of a column. expression A column name, a constant, a function, any combination of column names, constants, and functions connected by arithmetic or bitwise operators, or a subquery. In the context of the aggregate functions, an expression is usually a column name. (See "Expressions" for more information.) SUM Finds the sum of all values in the column. SUM can be used with numeric columns only. Null values are ignored. AVG Finds the average of the values in the column. AVG can be used with numeric columns only. Null values are ignored. COUNT Finds the number of non-null values in the column. When DISTINCT is specified, COUNT finds the number of unique non-null values. COUNT can be used with both numeric and character columns. Null values are ignored. COUNT(*) Finds the number of rows. COUNT(*) does not take any parameters and cannot be used with DISTINCT. All rows are counted, regardless of the presence of null values. MAX Finds the maximum value in the column. MAX can be used with numeric, character, and datetime columns. With character columns, MAX finds the value that is highest in the collating sequence. MAX ignores any null values. MAX cannot be used with bit columns. DISTINCT is not available since it is not meaningful with MAX. MIN Finds the minimum value in the column. MIN can be used with numeric, character, and datetime columns. With character columns, MIN finds the value that is lowest in the sort sequence. MIN ignores any null values. MIN cannot be used with bit columns. DISTINCT is not available since it is not meaningful with MIN. Comments Aggregate functions can be used in the select list or in the HAVING clause of a SELECT statement. They cannot be used in a WHERE clause. Aggregate functions, which calculate summary values from the non-null values in a particular column, can be applied to all rows in a table, in which case they produce a single value, called a scalar aggregate function. Alternatively, they can be applied to all rows that have the same value in a specified column or expression (with the GROUP BY and, optionally, the HAVING clause), in which case they produce a value for each group, called a vector aggregate function. The results of the aggregate functions are shown as new columns. You can nest a vector aggregate function inside a scalar aggregate function. For example: ╓┌─┌─────────────────────────────────────────┌─────────────┌─────────────────╖ ──────────────────────────────────────────────────────────────────────────── select type, avg(price), avg(avg(price)) from titles group by type type ------------ ------------ ------------ ──────────────────────────────────────────────────────────────────────────── ------------ ------------ ------------ UNDECIDED NULL NULL business 13.73 15.23 mod_cook 11.49 15.23 popular_comp 21.48 15.23 psychology 13.50 15.23 trad_cook 15.96 15.23 (6 rows affected) The GROUP BY clause applies to the vector aggregate function─in this case, AVG(price). The scalar aggregate function, AVG(AVG(price)), returns the average price of all books in the titles table. In standard SQL, all the columns in a select list that includes an aggregate function must either have aggregate functions applied to them or be in the GROUP BY list. TRANSACT-SQL has no such restrictions. Example A shows a SELECT statement with the standard restrictions; example B shows the same statement with another item (title_id) added to the select list to illustrate the difference in displays. These extra columns can also be referenced in a HAVING clause. A. select type, avg(price), avg(advance) from titles group by type ╓┌─┌──────────────────┌─────────────┌────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type ------------ ------------ ------------ UNDECIDED NULL NULL business 13.73 6,281.25 mod_cook 11.49 7,500.00 popular_comp 21.48 7,500.00 psychology 13.50 4,255.00 trad_cook 15.96 6,333.33 ──────────────────────────────────────────────────────────────────────────── trad_cook 15.96 6,333.33 (6 rows affected) B. select type, title_id, avg(price), avg(advance) from titles group by type order by type ╓┌─┌───────────────────┌───────────┌───────────┌─────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type title_id ------------ ---------- ---------- ----------- UNDECIDED MC3026 NULL NULL business BU1032 13.73 6,281.25 business BU1111 13.73 6,281.25 business BU2075 13.73 6,281.25 business BU7832 13.73 6,281.25 mod_cook MC2222 11.49 7,500.00 mod_cook MC3021 11.49 7,500.00 ──────────────────────────────────────────────────────────────────────────── mod_cook MC3021 11.49 7,500.00 popular_comp PC1035 21.48 7,500.00 popular_comp PC8888 21.48 7,500.00 popular_comp PC9999 21.48 7,500.00 psychology PS1372 13.50 4,255.00 psychology PS2091 13.50 4,255.00 psychology PS2106 13.50 4,255.00 psychology PS3333 13.50 4,255.00 psychology PS7777 13.50 4,255.00 trad_cook TC3218 15.96 6,333.33 trad_cook TC4203 15.96 6,333.33 trad_cook TC7777 15.96 6,333.33 (18 rows affected) You can use either a column name or any other expression (except a column heading or alias) after GROUP BY. Null values in the GROUP BY column are put into a single group. The COMPUTE clause in a SELECT statement uses row aggregate functions to produce summary values. The row aggregate functions make it possible to retrieve detail and summary rows with one statement. The following example illustrates this feature: select type, title_id, price, advance from titles where type = "psychology" order by type compute sum(price), sum(advance) by type ╓┌───────┌───────────────────┌─────────────┌───────────────────┌───────────── ───────────────────────────────────────────────────────────────────────────── type title_id price advance ------------ ------------ ---------- ----------- psychology PS1372 21.59 7,000.00 psychology PS2091 10.95 2,275.00 ───────────────────────────────────────────────────────────────────────────── psychology PS2091 10.95 2,275.00 psychology PS2106 7.00 6,000.00 psychology PS3333 19.99 2,000.00 psychology PS7777 7.99 4,000.00 sum sum -----------67.52 -------------21 (6 rows affected) Note the difference in display between this example and the earlier examples without COMPUTE. You can use a HAVING clause without a GROUP BY clause. Because each aggregate function in a query requires its own work table, a query using aggregate functions can exceed the maximum number of tables allowed in a query (16). If there are columns in the select list that neither have aggregate functions applied to them nor are included in the query's GROUP BY clause (illegal in standard SQL), the meanings of HAVING and WHERE are somewhat different. In this situation, a WHERE clause restricts the rows included in the calculation of the aggregate function but does not restrict the rows returned by the query. Conversely, a HAVING clause restricts the rows returned by the query but does not affect the calculation of the aggregate function. See "GROUP BY and HAVING Clauses" for examples. When you sum or average integer data, TRANSACT-SQL treats the result as an int value, even if the datatype of the column is smallint or tinyint. To avoid overflow errors in DB-LIBRARY programs, declare all variables for results of averages or sums as type int. See Also COMPUTE Clause, Functions, GROUP BY and HAVING Clauses, Row Aggregate Functions, Search Conditions, SELECT, WHERE Clause ALTER DATABASE ──────────────────────────────────────────────────────────────────────────── Function Increases the amount of disk space allocated to a database. Syntax ALTER DATABASE database_name [ON {DEFAULT | database_device} [=size] [, database_device [= size]]...] Examples A. alter database pubs B. alter database pubs on file1 = 3 Options database_name The name of the database whose storage size is to be changed. ON Indicates a size and/or location for the database extension. The default extension size is 1 megabyte. DEFAULT Indicates that ALTER DATABASE can put the database extension on any default database device(s) (as shown in sysdevices.status). To specify a size for the database extension without specifying the location, use "ON DEFAULT = size." (The default size is 1 megabyte.) To change a database device's status to the default, use the sp_diskdefault system procedure. database_device The logical name of the database device on which you want to locate the database extension. The default size of database_device is 2 megabytes. A database can occupy more than one database device with different amounts of space on each. size The amount of space allocated to the database extension, in megabytes. The minimum extension is 1 megabyte. The default value is 1 megabyte for a default database device and 2 megabytes for a nondefault database device. Legal values range from 1 to 215. If SQL Server can't allocate the requested size, it will do as much as it can in half-megabyte units with a minimum of 1 megabyte. Comments The Database Owner or System Administrator must be using the master database to execute the ALTER DATABASE statement. If you don't specify a location or size, the default location is any default database device indicated in sysdevices and the default size is 1 megabyte. If SQL Server can't handle your request, it allocates as much space as possible on each database device and prints a message telling how much space has been allocated on each database device. It is important to back up the master database with the DUMP DATABASE or DUMP TRANSACTION statement after each use of ALTER DATABASE. This makes recovery easier and safer if the master database is damaged. (If you use ALTER DATABASE and fail to back up the master database, you may subsequently be able to recover the changes with DISK REFIT.) To increase the amount of storage space allocated for the transaction log, give the name of the log's database device in the ON clause when you execute the ALTER DATABASE statement. Then execute sp_logdevice to make the newly allocated space available for the log. To rename a database, execute the sp_renamedb system procedure, which takes the old name of the database and its new name as parameters. For information on a database, execute sp_helpdb. To get a report on the amount of space used in a database, execute sp_spaceused. Permissions ALTER DATABASE permission defaults to the Database Owner if he or she has CREATE DATABASE permission. It cannot be transferred. This prevents unauthorized users from allocating too much disk space. The System Administrator can also alter databases. See Also CREATE DATABASE, DROP DATABASE, sp_diskdefault, sp_helpdb, sp_logdevice, sp_renamedb, sp_spaceused ALTER TABLE ──────────────────────────────────────────────────────────────────────────── Function Adds new columns to an existing table. Syntax ALTER TABLE [[database.]owner.]table_name ADD column_name datatype NULL [, column_name datatype NULL...] Examples alter table publishers add manager_name varchar(40) null Options datatype Any of the system datatypes except bit and any user-defined datatype except those based on bit. NULL A required parameter because the initial values in the new column must be set to NULL in already existing rows. Comments Columns of type bit cannot be added to an existing table. The number of columns in a table cannot exceed 250. To rename a table, execute the sp_rename system procedure. For information on a table and its columns, use sp_help. Permissions ALTER TABLE permission defaults to the table owner. It cannot be transferred. The System Administrator can alter users' tables. See Also CREATE TABLE, DROP TABLE, sp_help, sp_rename Batch Queries ──────────────────────────────────────────────────────────────────────────── Function A batch or batch file is a set of SQL statements submitted together and executed as a group, one after the other. A batch is terminated by an end-of-batch signal. With the isql program, the signal is the go command on a line by itself. For details on the isql program, see Chapter 3, "Utility Programs." Examples A. select count(*) from titles select count(*) from authors go B. create table test (column1 char(10), column2 int) insert test values ("hello", 598) select * from test go C. use master go select count(*) from sysdatabases go Comments CREATE PROCEDURE, CREATE RULE, CREATE DEFAULT, CREATE TRIGGER, and CREATE VIEW statements cannot be combined with other statements in a batch. They must be submitted singularly. CREATE DATABASE, CREATE TABLE, and CREATE INDEX can be combined with other statements in a single batch. USE must be submitted in a batch before you can reference objects in that database. You cannot drop an object and then reference or re-create it in the same batch. Any options set with a SET statement take effect at the end of the batch. You can combine SET statements and queries in the same batch, but the SET options won't apply to the queries in that batch. You can submit batches through the SQL Server Administration Facility or from a file with the isql program. A file submitted to isql can include more than one batch of SQL statements if each batch is terminated by the go command. If there is an error anywhere in a batch, none of its statements are executed. A batch is considered a transaction. See Also CREATE DEFAULT, CREATE PROCEDURE, CREATE RULE, CREATE TRIGGER, CREATE VIEW, SET, USE BEGIN...END ──────────────────────────────────────────────────────────────────────────── Function Enclose a series of SQL statements so that control-of-flow language, such as IF...ELSE, affects the performance of the whole group. Syntax BEGIN statement_block END Examples A. if (select avg(price) from titles) < $15 begin update titles set price = price * $2 select title, price from titles where price > $28 end Without BEGIN and END, the IF condition would cause execution of only one SQL statement. B. create trigger deltitle on titles for delete as if (select count(*) from deleted, sales where sales.title_id = deleted.title_id) > 0 begin rollback transaction print "You can't delete a title with sales." end else print "Deletion successful─no sales for this title." Options statement_block A series of statements enclosed by BEGIN and END. Comments BEGIN...END blocks can be nested within other BEGIN...END blocks. See Also Control-of-Flow Language BEGIN TRANSACTION ──────────────────────────────────────────────────────────────────────────── Function Marks the starting point of a user-specified transaction. Syntax BEGIN TRANsaction [transaction_name] Examples begin transaction royaltychange /* A user sets out to change the royalty split for the two authors of The Gourmet Microwave. Since the database would be inconsistent between the two updates, they must be grouped into a user-defined transaction. */ update titleauthor set royaltyper = 65 from titleauthor, titles where royaltyper = 75 and titleauthor.title_id = titles.title_id and title = "The Gourmet Microwave" update titleauthor set royaltyper = 35 from titleauthor, titles where royaltyper = 25 and titleauthor.title_id = titles.title_id and title = "The Gourmet Microwave" save transaction percentchanged /* After having updated the royaltyper entries for the two authors, the user inserts the savepoint percentchanged and then determines how a 10% increase in the book's price would affect the authors' royalty earnings. */ update titles set price = price * $1.1 where title = "The Gourmet Microwave" select (price * royalty * ytd_sales) * royaltyper from titles, titleauthor where title = "The Gourmet Microwave" and titles.title_id = titleauthor.title_id /* The transaction is rolled back to the savepoint with the ROLLBACK TRANSACTION statement. */ rollback transaction percentchanged commit transaction /* End of royaltychange. */ Options transaction_name The name assigned to this transaction. It must conform to the rules for identifiers. Comments Define a transaction by enclosing SQL statements and/or stored procedures within the statements BEGIN TRANSACTION and COMMIT TRANSACTION. To cancel an entire transaction, use the statement ROLLBACK TRANSACTION transaction_name. All of the transaction's statements or procedures that have been completed are undone. To cancel part of a transaction, use the statement ROLLBACK TRANSACTION savepoint_name. (The savepoint was inserted inside the transaction earlier with a SAVE TRANSACTION statement.) All of the transaction's statements or procedures between the savepoint and the ROLLBACK TRANSACTION statement are undone. After a transaction is rolled back to a savepoint, it must proceed to completion (with more SQL statements if desired and a COMMIT TRANSACTION statement), or it must be canceled altogether (by rolling it back to its beginning). The ROLLBACK TRANSACTION statement must appear within a transaction: you can't rollback a transaction after COMMIT TRANSACTION has been entered. If no savepoint_name or transaction_name is given with the ROLLBACK TRANSACTION statement, the transaction is rolled back to the previous BEGIN TRANSACTION. The following statements cannot be used inside a user-defined transaction: CREATE DATABASE, CREATE TABLE, CREATE INDEX, all DROP statements, SELECT INTO (because it creates a table), GRANT, REVOKE, ALTER DATABASE, ALTER TABLE, TRUNCATE TABLE, RECONFIGURE, LOAD DATABASE, LOAD TRANSACTION, and DISK INIT. In addition, some of the system procedures cannot be used inside user-defined transactions because they create temporary tables. Transactions can be nested inside each other. Permissions BEGIN TRANSACTION permission defaults to all users. No permission is required to use it. See Also COMMIT TRANSACTION, ROLLBACK TRANSACTION, SAVE TRANSACTION BREAK ──────────────────────────────────────────────────────────────────────────── Function Controls operation of statements in a WHILE loop. BREAK is often (but not always) activated by an IF test. BREAK causes an exit from the WHILE loop. Syntax WHILE boolean expression statement BREAK statement CONTINUE Examples while (select avg(price) from titles) < $30 begin update titles set price = price * $2 select max(price) from titles if (select max(price) from titles) > $50 break else continue end begin print "Too much for the market to bear" end If the average price is less than $30, double the prices. Then select the maximum price. If it is less than or equal to $50, restart the WHILE loop and double the prices again. If the maximum price is more than $50, exit the WHILE loop and print a message. Comments BREAK causes an exit from the WHILE loop. Any statements that appear after the END keyword that marks the end of the loop are executed. If two or more WHILE loops are nested, the inner BREAK exits to the next outermost loop. First, all the statements after the end of the inner loop run, and then the next outermost loop restarts. See Also CONTINUE, Control-of-Flow Language, Expressions, WHILE Browse Mode ──────────────────────────────────────────────────────────────────────────── Function Supports the ability to perform updates while viewing data. Browse mode is used in an application program using DB-LIBRARY. The FOR BROWSE clause is used in a SELECT statement. Syntax FOR BROWSE Examples See the SQL Server Programmer's Reference for an example of a DB-LIBRARY application that uses browse mode. Comments A table can be browsed in an application program if ■ Its rows have been timestamped by including a column named timestamp in the table definition or by adding the timestamp column with ALTER TABLE: CREATE TABLE browsetest (col1 int, timestamp, col3 char(7)) or ALTER TABLE table_name ADD timestamp A column named timestamp automatically has the system datatype timestamp and is automatically updated. ■ It has a unique key. ■ The keywords FOR BROWSE are at the end of the SELECT statement sent to SQL Server. The use of the keyword HOLDLOCK is forbidden in a SELECT statement that includes the FOR BROWSE clause. See Also Datatypes, SELECT, sp_primarykey CHECKPOINT ──────────────────────────────────────────────────────────────────────────── Function Forces all dirty pages (pages on which there have been updates since the last checkpoint) in the current database to be written to the disk. Checkpoints caused by the CHECKPOINT statement supplement automatic checkpoints, which occur at intervals calculated by SQL Server on the basis of the configurable value for maximum acceptable recovery time. Syntax CHECKPOINT Examples checkpoint All dirty pages in the current database are written to the disk, regardless of the system checkpoint schedule. Comments The CHECKPOINT statement saves work in recovery by identifying a point at which all completed transactions are guaranteed to have been written to the disk. A typical checkpoint takes one second, though this figure varies depending on the amount of activity on SQL Server. The automatic checkpoint interval is calculated by SQL Server on the basis of system activity and the recovery interval value in the system table syscurconfigs. The recovery interval determines checkpoint frequency by specifying the amount of time it should take the system to recover. Reset this value by executing the sp_configure system procedure and the RECONFIGURE statement. For more information, see the SQL Server System Administrator's Guide. Use the CHECKPOINT statement as a precautionary measure only in special circumstances. For example, if the automatic checkpoint is infrequent because the recovery interval is set high, you might want to use the CHECKPOINT statement after inserting rows with bulk copy. Use the CHECKPOINT statement after changing a database option with the sp_dboption system procedure. If you change a database option with sp_dboption inside a user-defined transaction and then roll back that transaction you must use another CHECKPOINT statement to make the rollback take effect on the option change. Here's an example: begin transaction use master <execute> sp_dboption 'pubs', 'single', 'true' <execute> use pubs <execute> checkpoint <execute> rollback transaction <execute> /* ** If the following checkpoint is not used, the ** "pubs" database remains single-user. */ checkpoint <execute> Permissions Permission to use the CHECKPOINT statement defaults to the Database Owner. It cannot be transferred. See Also RECONFIGURE, sp_dboption Comments ──────────────────────────────────────────────────────────────────────────── Function Provide information about SQL statements, statement blocks, and stored procedures. Syntax /* text of comment */ Examples /* this procedure finds rules by user name */ create procedure findmyrule @nm varchar(30) = null as if @nm is null begin print "You must give a user name" return print "I have returned" /* this statement follows RETURN, ** so won't be executed */ end else /* print the rule names and IDs, and the user ID */ select sysobjects.name, sysobjects.id, sysobjects.uid from sysobjects, master..syslogins where master..syslogins.name = @nm and sysobjects.uid = master..syslogins.suid and sysobjects.type = "R" Comments Comments can be inserted on a line by themselves or at the end of a command line. Multiple-line comments are allowed as long as they are surrounded by /* and */. A stylistic convention often used for multiple-line comments is to begin the first line with /* and subsequent lines with **. The comment is ended with */. There is no maximum length for comments. Comments can be nested. ──────────────────────────────────────────────────────────────────────────── NOTE Do not include a go command within a comment because it will produce an error message. ──────────────────────────────────────────────────────────────────────────── COMMIT TRANSACTION ──────────────────────────────────────────────────────────────────────────── Function Marks the ending point of a user-defined transaction. Syntax COMMIT TRANsaction [transaction_name] Examples begin transaction royaltychange /* A user sets out to change the royalty split for the two authors of The Gourmet Microwave. Since the database would be inconsistent between the two updates, they must be grouped into a user-defined transaction. */ update titleauthor set royaltyper = 65 from titleauthor, titles where royaltyper = 75 and titleauthor.title_id = titles.title_id and title = "The Gourmet Microwave" update titleauthor set royaltyper = 35 from titleauthor, titles where royaltyper = 25 and titleauthor.title_id = titles.title_id and title = "The Gourmet Microwave" save transaction percentchanged /* After having updated the royaltyper entries for the two authors, the user inserts the savepoint percentchanged and then experiments to see how a 10% increase in the book's price would affect the authors' royalty earnings. */ update titles set price = price * $1.1 where title = "The Gourmet Microwave" select (price * royalty * ytd_sales) * royaltyper from titles, titleauthor where title = "The Gourmet Microwave" and titles.title_id = titleauthor.title_id /* The transaction is rolled back to the savepoint with the ROLLBACK TRANSACTION statement. */ rollback transaction percentchanged commit transaction /* End of royaltychange. */ Options transaction_name The name assigned to the transaction. It must conform to the rules for identifiers. (See "Identifiers" for details.) Comments Define a transaction by enclosing SQL statements and/or stored procedures with the BEGIN TRANSACTION and COMMIT TRANSACTION statements. A savepoint is a marker set by the SAVE TRANSACTION savepoint_name statement within a transaction. You can cancel part of a transaction with the following statement: rollback tran savepoint_name All of the statements or procedures between the savepoint and ROLLBACK TRANSACTION are undone. To cancel an entire transaction, use the ROLLBACK TRANSACTION statement. All of the transaction's statements or procedures are undone. If no savepoint name or transaction name is given with the ROLLBACK TRANSACTION statement, the transaction is rolled back to the previous BEGIN TRANSACTION. The ROLLBACK TRANSACTION statement must appear within a transaction: you can't rollback a transaction after COMMIT TRANSACTION has been entered. After a transaction is rolled back to a savepoint, it must proceed to completion (with more SQL statements if desired and a COMMIT TRANSACTION statement), or it must be canceled altogether (by rolling it back to its beginning). Until you execute COMMIT TRANSACTION, SQL Server will consider all subsequent statements to be part of the transaction, until it encounters another BEGIN TRANSACTION statement. At that point, SQL Server rolls the original transaction back and displays an error message. You may want to insert a comment giving the name of the transaction being ended, as in the example. Permissions COMMIT TRANSACTION permission defaults to all users. No permission is required to use it. See Also BEGIN TRANSACTION, ROLLBACK TRANSACTION, SAVE TRANSACTION COMPUTE Clause ──────────────────────────────────────────────────────────────────────────── Function Generates summary values in a SELECT statement with row aggregate functions (SUM, AVG, MIN, MAX, and COUNT). The summary values appear as additional rows in the query results (unlike the aggregate function results, which appear as new columns), which allows you to see the detail and summary rows in one set of results. You can calculate summary values for subgroups, and you can calculate more than one aggregate function for the same group. Syntax COMPUTE row_aggregate(column_name) [, row_aggregate(column_name)...] [BY column_name [, column_name]...] Examples A. select type, price from titles where price > $12 and type like "%cook" order by type, price compute sum(price) by type ╓┌─┌──────────────────┌──────────────────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type price --------- ------------ mod_cook 19.99 sum ------------ 19.99 type price --------- ------------ trad_cook 14.99 trad_cook 20.95 sum ------------ ──────────────────────────────────────────────────────────────────────────── ------------ 35.94 (5 rows affected) Example A calculates the sum of the prices of each type of cookbook that costs more than $12. B. select type, price, advance from titles titles where price > $12 and type like "%cook" order by type, price compute sum(price), sum(advance) by type ╓┌─┌──────────────────┌──────────┌───────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type price advance --------- --------- ------------ ──────────────────────────────────────────────────────────────────────────── --------- --------- ------------ mod_cook 19.99 0.00 sum sum --------- ------------ 19.99 0.00 type price advance --------- --------- ------------ trad_cook 14.99 8,000.00 trad_cook 20.95 7,000.00 sum sum --------- ------------ 35.94 15,000.00 (5 rows affected) Example B calculates the sum of the prices and advances for each type of cookbook that costs more than $12. C. select type, price, advance from titles where price > $12 and type like "%cook" order by type, price compute sum(price), max(advance) by type ╓┌─┌──────────────────┌─────────────┌────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type price advance --------- ------------ ------------ mod_cook 19.99 0.00 sum ------------ 19.99 max ------------ 0.00 type price advance ---------- ------------ ------------ trad_cook 14.99 8,000.00 trad_cook 20.95 7,000.00 ──────────────────────────────────────────────────────────────────────────── trad_cook 20.95 7,000.00 sum ------------ 35.94 max ------------ 8,000.00 (5 rows affected) Example C calculates the sum of the prices and maximum advance of each type of cookbook that costs more than $12. D. select type, pub_id, price from titles where price > $10 and type = "psychology" order by type, pub_id, price compute sum(price) by type, pub_id ╓┌─┌──────────────────┌────────────┌─────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type pub_id price ------------ --------- ------------ psychology 0736 10.95 psychology 0736 19.99 sum ------------ 30.94 type pub_id price ------------ ---------- ------------ psychology 0877 21.59 sum ------------ 21.59 (5 rows affected) Example D breaks on type and pub_id and calculates the sum of the prices of psychology books by type-publisher ID combination. E. select type, pub_id, price from titles where price > $10 and type = "psychology" order by type, pub_id, price compute sum(price) by type, pub_id compute sum(price) by type ╓┌─┌──────────────────┌───────────┌──────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type pub_id price ----------- --------- ------------ psychology 0736 10.95 psychology 0736 19.99 sum ------------ 30.94 type pub_id price ---------- --------- ------------ psychology 0877 21.59 sum ──────────────────────────────────────────────────────────────────────────── sum ------------ 21.59 sum ------------ 52.53 (6 rows affected) Example E calculates the grand total of the prices of psychology books that cost more than $10, in addition to sums by type and pub_id. F. select type, price, advance from titles where price > $10 and type like "%cook" compute sum(price), sum(advance) ╓┌─┌──────────────────┌─────────────┌────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── ──────────────────────────────────────────────────────────────────────────── type price advance --------- ----------- ------------ mod_cook 19.99 0.00 trad_cook 20.95 8,000.00 trad_cook 11.95 4,000.00 trad_cook 14.99 7,000.00 sum sum ------------ ----------- 67.88 19,000.00 (5 rows affected) Example F calculates the grand totals of the prices and advances of cooking books that cost over $10. Options row_aggregate One of the following: ╓┌─────────────────────────────────┌─────────────────────────────────────────╖ Function Definition ──────────────────────────────────────────────────────────────────────────── SUM Total of values in the (numeric) column AVG Average of values in the (numeric) column MIN Lowest value in the column MAX Highest value in the column COUNT Number of values in the column column_name The name of a column, which must be enclosed in parentheses. Only numeric columns can be used with SUM and AVG. One COMPUTE clause can apply the same function to several columns. (See example B.) When using more than one Function Definition ──────────────────────────────────────────────────────────────────────────── example B.) When using more than one function, use more than one COMPUTE clause. (See example E.) BY Indicates that row aggregate function values are to be calculated for subgroups. Whenever the value of the BY item changes, row aggregate function values are generated. If you use BY, you must use ORDER BY. Listing more than one item after BY breaks a group into subgroups and applies a function at each level of grouping. ──────────────────────────────────────────────────────────────────────────── Comments Row aggregate functions make it possible to retrieve detail and summary rows with one statement. Aggregate functions, on the other hand, ordinarily produce a single value for all selected rows in the table or for each group, and these summary values are shown as new columns. The following examples illustrate the differences: A. select type, sum(price), sum(advance) from titles where type like "%cook" group by type ──────────────────────────────────────────────────────────────────────────── type ------------ --------- ---------- mod_cook 22.98 15,000.00 trad_cook 47.89 19,000.00 (2 rows affected) B. select type, price, advance from titles where type like "%cook" order by type compute sum(price), sum(advance) by type ╓┌─┌──────────────────┌─────────────┌────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type price advance --------- ------------ ------------ mod_cook 2.99 15,000.00 mod_cook 19.99 0.00 sum sum ------------ ------------ 22.98 15,000.00 type price advance --------- ------------ ------------ trad_cook 11.95 4,000.00 trad_cook 14.99 8,000.00 trad_cook 20.95 7,000.00 sum sum ----------- ------------ ──────────────────────────────────────────────────────────────────────────── ----------- ------------ 47.89 19,000.00 (7 rows affected) The columns in the COMPUTE clause must appear in the select list. You can't use SELECT INTO in the same statement with a COMPUTE clause because statements that include COMPUTE do not generate normal tables. If you use COMPUTE BY, you must also use an ORDER BY clause. The columns listed after COMPUTE BY must be identical to or a subset of those listed after ORDER BY, and must be in the same left-to-right order, start with the same expression, and not skip any expressions. For example, if the ORDER BY clause is order by a, b, c the COMPUTE BY clause can be any of these: compute by a, b, c compute by a, b compute by a The COMPUTE keyword can be used without BY to generate grand totals, grand counts, and so on. ORDER BY is optional if you use the COMPUTE keyword without BY. See example F. In a SELECT statement with a COMPUTE clause, the order of columns in the select list overrides the order of the aggregate functions in the COMPUTE clause. DB-LIBRARY programmers must be aware of this to put the aggregate function results in the right place. See "Row Aggregate Functions" for an example. See Also Aggregate Functions, GROUP BY and HAVING Clauses, Row Aggregate Functions, SELECT CONTINUE ──────────────────────────────────────────────────────────────────────────── Function Controls operation of statements in a WHILE loop. CONTINUE is often (but not always) activated by an IF test. CONTINUE causes the WHILE loop to restart. Syntax WHILE boolean expression statement BREAK statement CONTINUE Examples while (select avg(price) from titles) < $30 begin update titles set price = price * $2 select max(price) from titles if (select max(price) from titles) > $50 break else continue end begin print "Too much for the market to bear" end If the average price is less than $30, double the prices. Then select the maximum price. If it is less than or equal to $50, restart the WHILE loop and double the prices again. If the maximum price is more than $50, exit the WHILE loop and print a message. Comments CONTINUE causes the WHILE loop to restart, skipping any statements after CONTINUE. See Also BREAK, Expressions, WHILE Control-of-Flow Language ──────────────────────────────────────────────────────────────────────────── Function Controls the flow of execution of SQL statements, statement blocks, and stored procedures. Statements BEGIN Begins a statement block. ...END Ends a statement block. BREAK Exits the innermost WHILE loop. ...CONTINUE Restarts a WHILE loop. DECLARE Declares local variables. GOTO label: Goes to label:, a position in a statement block. IF Defines conditional execution. ...ELSE Defines alternate execution when a condition is false. PRINT Prints a user-defined message on the user's screen. RAISERROR Prints a user-defined message on the user's screen and sets a system flag (in the global variable @ERROR) to record the fact that an error condition has occurred. RETURN Exits unconditionally. WAITFOR Sets a delay for statement execution. WHILE Repeats statements while condition is true. /* comment */ Inserts a comment anywhere in an SQL statement. Comments Control-of-flow language can be used with interactive SQL statements, in batches, and in stored procedures. See Also Batch Queries, BEGIN...END, BREAK, Comments, CONTINUE, CREATE PROCEDURE, CREATE TRIGGER, DECLARE, GOTO, IF...ELSE, Parameters, PRINT, RAISERROR, RETURN, Variables (Local and Global), WAITFOR, WHILE Conversion Function ──────────────────────────────────────────────────────────────────────────── Function Converts expressions of one datatype to another datatype. Also obtains a variety of special date formats. Syntax CONVERT(datatype, expression [, style]) Examples A. select title, convert(char(12), ytd_sales) from titles B. select title, ytd_sales from titles where convert(char(20), ytd_sales) like "1%" C. select convert(char(12), getdate(), 3) Example C converts the current date to style "3", dd/mm/yy. Options datatype Any system datatype (for example, char(10), varbinary(50), int) into which the expression is to be converted. User-defined datatypes cannot be used. If no length is specified when converting to char, varchar, binary, or varbinary datatypes, the data adapts to any size necessary. The maximum length you can specify when converting to these types is 255. expression The value to be converted. style The style of date representation you desire when converting to char or varchar from datetime data. Add 100 to a style value to get a four-place year that includes the century (yyyy). ╓┌───────────────────┌──────────────────┌──────────────────┌───────────────── Without With Standard Output Century Century (yy) (yyyy) ───────────────────────────────────────────────────────────────────────────── - 0 or 100 default mon dd yyyy hh:miAM (or PM) 1 101 USA mm/dd/yy 2 102 ANSI yy.mm.dd 3 103 British/French dd/mm/yy 4 104 German dd.mm.yy 5 105 Italian dd-mm-yy 6 106 - dd mon yy Without With Standard Output Century Century (yy) (yyyy) ───────────────────────────────────────────────────────────────────────────── 6 106 - dd mon yy 7 107 - mon dd yy 8 108 - hh:mi:ss - 9 or 109 default + mon dd, yyyy milliseconds hh:mi:sssAM (or PM) 10 110 USA mm-dd-yy 11 111 JAPAN yy/mm/dd 12 112 ISO yymmdd ───────────────────────────────────────────────────────────────────────────── Without With Standard Output Century Century (yy) (yyyy) ───────────────────────────────────────────────────────────────────────────── ───────────────────────────────────────────────────────────────────────────── The default values (style 0 or 100 and 9 or 109) always return the century (yyyy). Comments Table shows the possibilities for converting datatypes. Table 1.1 Datatype Conversion Chart ╓┌───────────────┌───────┌──────────┌──────────────┌───────────────┌───────── ───────────────────────────────────────────────────────────────────────────── ───────────────────────────────────────────────────────────────────────────── To: binary varbinary tinyint(INT1) smallint(INT2) int(INT4) From: binary - I I I I varbinary I - I I I tinyint(INT1) E I - I I smallint(INT2) E I I - I int(INT4) E E I I - float N N I I I char E E E E E varchar E E E E E ───────────────────────────────────────────────────────────────────────────── money I I I I I bit I I I I I datetime E E N N N text N N N N N image E E N N N ───────────────────────────────────────────────────────────────────────────── Key: I Implicit conversion E Explicit conversion, CONVERT function must be used N Conversion not allowed - Conversion of a datatype to itself; allowed but meaningless SQL Server automatically handles certain datatype conversions. For example, if you compare a char expression and a datetime expression, or a smallint expression and an int expression, or char expressions of different lengths, SQL Server makes the conversion automatically for the comparison. You need not use the CONVERT function. It is never wrong to use the CONVERT function, even when you are comparing two expressions of exactly the same datatype. Automatic conversion is not supported for the text and image datatypes. You can explicitly convert text data to char or varchar, and image data to binary or varbinary, but the maximum length you can specify is 255. If you attempt a conversion that is not possible (for example, if you try to convert a char expression that includes letters to an int), SQL Server generates an error message. If you do not specify a length for the char, varchar, binary, or varbinary datatype to which the expression is to be converted, the data adapts to any size necessary. When converting to datetime, SQL Server rejects all values it cannot recognize as dates (including dates earlier than January 1, 1753). Converting to bit changes any non-zero value to a 1. When converting to money from integer datatypes, units are assumed to be dollars. For example, the integer value of 4 would be converted to the money equivalent of 4 dollars. Expressions of type char or varchar being converted to an integer datatype must consist only of digits and an optional plus (+) or minus (-) sign. Leading blanks are ignored. Expressions of type char or varchar being converted to money can also include an optional decimal point and dollar sign ($). Expressions of type char or varchar being converted to float can also include optional exponential notation (a lowercase "e" or uppercase "E" followed by an optional + or - and then a number). When char or varchar expressions are converted to a character datatype of a different size, values too long for the new datatype are truncated. Converting to binary is used to display the internal representation of a value. If lengths differ, the binary value is padded with zeros on the right. You can convert int, smallint, and tinyint to binary or varbinary, but if you convert the binary value back to an integer value, it will be different. In addition, if the integer is longer than the specified binary length, it is truncated without notice. When converting between types with a different number of decimal points, the value is truncated. For example, the result of the following is "10": select convert(int, 10.672) You can convert any datatype to binary or varbinary, but if the binary field is too short and the value being converted is truncated, the data may be meaningless. The type conversion function can be used in the select list, in the WHERE clause, and anywhere an expression is allowed. See Also Datatypes, Date Functions, Functions, Mathematical Functions, String Functions, Text/Image Functions CREATE DATABASE ──────────────────────────────────────────────────────────────────────────── Function Creates a new database. You must be in the master database to create a new database. Syntax CREATE DATABASE database_name [ON {DEFAULT | database_device} [= size] [, database_device [= size]]...] Examples A. create database pubs B. create database pubs on default = 4 C. create database pubs on file1 = 3, file2 = 2 Options database_name The name of the new database. It must conform to the rules for identifiers. ON Indicates that you want to specify a location and (optionally) a size for the database. DEFAULT Indicates that CREATE DATABASE can put the new database on any default database device(s) (as shown in sysdevices.status). To specify a size for the database without specifying a location, use "ON DEFAULT = size". To change a database device's status to the default, use the sp_diskdefault system procedure. database_device The logical name of the database device on which you want to locate the database. A database can occupy different amounts of space on each of several database devices. DISK INIT is used to allocate space for a database device. size The amount of space allocated to the database, in megabytes. The SQL-Server-supplied default size is 2 megabytes. The System Administrator can change the default size by editing the system table master..sysconfigures, executing the RECONFIGURE statement and restarting SQL Server. The default size is usually the same size as the model database and must be at least as large as the model database. The legal values for a database size range from 2 to 215. Comments You must be using the master database to create a new user database. SQL Server can manage up to 32,767 databases. When the CREATE DATABASE statement is executed, SQL Server makes a copy of the model database, which initially contains the system tables. You can update the model database just like any other database to add your own tables, stored procedures, user datatypes, system variables, and so on. The default database size in master..sysconfigures must be at least as large as the model database. Only one database can be created at a time. If two database creation requests collide, one user will get the message "MODEL database in use: cannot create new database." If you don't specify a location and size for a database, the default location is any default database device indicated in master..sysdevices, and the default size is the size of the model database. If SQL Server can't give you as much space as you want where you've requested it, it creates the database anyway, giving you as much space as possible on each database device and prints a message telling how much space was allocated where. Each new database inherits its database option settings from the model database. For example, the database option select into/bulkcopy is originally set to "off". You can change that default with the sp_dboption system procedure. Since the CREATE DATABASE statement must clear every page (to guarantee the database can be recovered), it takes some number of seconds per megabyte to complete. It is important to back up the master database with the DUMP DATABASE or DUMP TRANSACTION statement after each use of CREATE DATABASE. This makes recovery easier and safer in case the master database is damaged. (If you create a database and fail to back up master, you may be able to recover the changes with DISK REFIT.) Whenever you create a database larger than about 4 megabytes, you should assign it to at least two separate database devices with the CREATE DATABASE statement and then use sp_logdevice to put the database's syslogs table on one of these database devices. This is essential to guarantee that SQL Server's roll-forward mechanism works correctly. Putting the transaction log and the rest of the database on different database devices also improves performance. ──────────────────────────────────────────────────────────────────────────── NOTE A smaller database can be created on a single database device and the transaction log stored together with the rest of the database, but you must rely on the DUMP DATABASE statement for backups. ──────────────────────────────────────────────────────────────────────────── The size of the database device required for the transaction log varies according to the amount of update activity and the frequency of transaction log dumps. As a rule, allocate to the log database device 10% to 25% of the space you allocate to the database itself. It is best to start small, since space allocated to a transaction log database device cannot be reclaimed and cannot be used for storage of data. To display a report on a database or on all databases on SQL Server, execute the sp_helpdb system procedure. For a report on the space used in a database, use sp_spaceused. Permissions CREATE DATABASE permission defaults to the System Administrator, who can transfer it to other users who are listed in the sysusers table of the current database. However, CREATE DATABASE permission is often centralized to maintain control over disk allocation. See the SQL Server System Administrator's Guide for details. CREATE DATABASE permission is not included when you use the GRANT ALL statement. See Also ALTER DATABASE, DISK INIT, DROP DATABASE, sp_changedbowner, sp_diskdefault, sp_helpdb, sp_logdevice, sp_renamedb, sp_spaceused CREATE DEFAULT ──────────────────────────────────────────────────────────────────────────── Function Specifies a value that will be inserted in a column (or in all columns of a user-defined datatype) if no value is explicitly supplied at insert time. Syntax CREATE DEFAULT [owner.]default_name AS constant_expression Examples create default namedefault as "unknown" A default value has been defined. Now you need to bind it to the appropriate column or user datatype with the sp_bindefault system procedure. Options default_name The name of the default. It must conform to the rules for identifiers. constant_expression An expression that contains only constant values (that is, it does not include the names of any columns or other database objects). Built-in functions that do not reference database objects can be included. Comments After you've created a default, use the sp_bindefault system procedure to bind it to a column or user-defined datatype. Defaults cannot be bound to SQL-Server-supplied datatypes. The default must be compatible with the datatype of the column. You cannot use "N/A", for example, as a default for a numeric column. If the default is not compatible with the column to which you've bound it, SQL Server will generate an error message when it tries to insert the default value (not when you bind it). Be sure to enclose character and date constants in quotation marks and to precede binary constants with "0x". If the default value is too long for the column to which you've bound it, the value will be truncated. CREATE DEFAULT statements cannot be combined with other SQL statements in a single batch. You can create a default in the current database only. Default definitions are stored in syscomments. After a default is bound to a particular column or user datatype, its object ID number is stored in the syscolumns or systypes system table. To display definitions and binding information, execute the sp_helptext system procedure. You must drop a default before you create a new one of the same name and you must unbind a default (with the sp_unbindefault system procedure) before you drop it. You can bind a new default to a column or datatype without unbinding the old one; the new default overrides the old one. You can bind one default to a user datatype and another default to a column of that datatype. The default most recently bound takes precedence. If a column has both a default and a rule associated with it, the default value must not violate the rule. A default that conflicts with a rule will never be inserted. SQL Server will generate an error message each time it attempts to insert the default. If you specify NOT NULL when you create a column and do not create a default for it, an error message will be generated whenever a user fails to make an entry in that column. The following table illustrates the relationship between the existence of a default and the definition of a column as NULL or NOT NULL. The entries in the table show the result: No entry ──────────────────────────────────────────────────────────────────────────── NULL null default null null NOT error default error error NULL ──────────────────────────────────────────────────────────────────────────── You can find the value of a default by using the sp_helptext system procedure with the default name as the parameter. To rename a default, use sp_rename. For a report on a default, use sp_help. Permissions CREATE DEFAULT permission defaults to the Database Owner, who can transfer it to other users. See Also Batch Queries, CREATE RULE, DROP DEFAULT, DROP RULE, sp_bindefault, sp_help, sp_helptext, sp_rename, sp_unbindefault CREATE INDEX ──────────────────────────────────────────────────────────────────────────── Function Creates indexes. Syntax CREATE [UNIQUE] [CLUSTERED | NONCLUSTERED] INDEX index_name ON [[database.]owner.]table_name (column_name [, column_name]...) [WITH {FILLFACTOR = x , IGNORE_DUP_KEY , [IGNORE_DUP_ROW | ALLOW_DUP_ROW]}] Examples A. create index au_id_ind on authors (au_id) B. create unique clustered index au_id_ind on authors (au_id) C. create index ind1 on titleauthor (au_id, title_id) D. create nonclustered index zip_ind on authors (zip) with fillfactor = 25 Options UNIQUE Prohibits duplicate index, or key, values. The system checks for duplicate key values when the index is created (if data already exists) and checks each time data is added with an INSERT or UPDATE. You cannot create a unique index if there are duplicate key values: the statement is aborted and an error message giving the first duplicate is printed. UPDATE or INSERT statements that would generate duplicate key values are rolled back, and SQL Server displays an error message. This is true even if the UPDATE or INSERT statement would have changed many rows but caused only one duplicate. Composite indexes (indexes in which the key value is composed of more than one column) can be unique. CLUSTERED Indicates that the physical order of rows is the same as the indexed order of the rows, and the bottom, or leaf, level of the clustered index contains the actual data pages. A clustered index almost always retrieves data faster than a nonclustered index. By definition, only one clustered index is permitted per table. The clustered index is often created on the table's primary key─the column or columns that uniquely identify the row. The primary key is declared with the sp_primarykey system procedure. NONCLUSTERED Indicates that there is an extra level between the index structure and the data itself. You can have up to 250 nonclustered indexes per table; however, you cannot exceed a total of 250 indexes, both clustered and nonclustered. If CLUSTERED is not specified, NONCLUSTERED is assumed. index_name The name of the index. Index names must be unique within a table but need not be unique within a database. table_name The name of the table in which the indexed column or columns are located. column_name The column or columns to which the index applies. Composite indexes are based on the combined values in up to 16 columns. The sum of the maximum lengths of all the columns used in a composite index cannot exceed 256 bytes. List the columns to be included in the composite index (in sort-priority order) inside the parentheses after table_name. FILLFACTOR Specifies how full SQL Server will make each page when it is creating a new index on existing data. The fillfactor percentage affects performance, because SQL Server must take the time to split pages when they fill up. The fillfactor percentage is relevant only at the time the index is created and becomes less important as the data changes. The pages are not maintained at any particular level of fullness. The default for FILLFACTOR is 0. Legal values are between 0 and 100. A fillfactor of 0 does not mean that pages are 0% full. Rather, it is like 100 in that SQL Server creates clustered indexes with completely full pages and nonclustered indexes with completely full leaf pages. It is different from 100 in that SQL Server leaves a comfortable amount of space within the index B-tree in both the clustered and nonclustered case. There is seldom a reason to change the fillfactor variable, especially since you can override it in the CREATE INDEX statement. If FILLFACTOR is 100, SQL Server creates both clustered and nonclustered indexes with each page 100% full. A fillfactor of 100 makes sense only for read-only tables─tables to which no additional data will ever be added. Smaller fillfactor values (except 0, which is a special case) cause SQL Server to create new indexes with pages that are not completely full. For example, a fill factor of 10 might be a reasonable choice if you are creating an index on a table that you know contains a small portion of the data it will eventually hold. Smaller fillfactor values cause each index to take more storage space. The default FILLFACTOR value can be altered by updating sysconfigures. IGNORE_DUP_KEY An infrequently used keyword. Its most common function is to sort through a batch of data that includes duplicate rows you wish to eliminate. IGNORE_DUP_KEY controls what happens when you attempt to enter a duplicate key in a unique clustered index. It is meaningful only when the UPDATE or INSERT statement affects multiple rows. If IGNORE_DUP_KEY is not set and an attempt is made to enter data that would create duplicate keys, the entire data modification statement fails. (If there are other statements in the same transaction, they execute normally.) If IGNORE_DUP_KEY is set and you give an UPDATE or INSERT statement that would create duplicate keys, the row that would cause the duplicates is not added or changed. In fact, in the case of UPDATE, the row is discarded. (See the following warning.) Other changes to the database caused by the UPDATE or INSERT attempt (for example, changes to index pages) are also backed out. However, if the UPDATE or INSERT affects multiple rows, the other rows are added or changed as usual. ──────────────────────────────────────────────────────────────────────────── WARNING With IGNORE_DUP_KEY set, if you try to update a row in such a way that a duplicate key would be created, that row is discarded. Neither the new value nor the original value of the row that would have produced the duplicate exists in the updated table. For example, if you try to update "Smith" to "Jones" and "Jones" already exists, you will have one "Jones" and no "Smith." The disappearance of the row for "Smith" is distressing, but unavoidable, because an UPDATE statement is actually a DELETE followed by an INSERT. SQL Server has no way to know about the disallowed duplicate when it deletes the row, and the whole transaction can't be rolled back because the purpose of IGNORE_DUP_KEY (and of the IGNORE_DUP_ROW option) is to allow a transaction to proceed in spite of the presence of duplicates. ──────────────────────────────────────────────────────────────────────────── You cannot create a unique index on a column that includes duplicate values, whether or not IGNORE_DUP_KEY is set. If you attempt to do so, SQL Server prints an error message that gives the first of the duplicate values. You must eliminate duplicates before you create a unique index on the column. IGNORE_DUP_ROW and ALLOW_DUP_ROW Keywords allowed only with a nonunique clustered index. (From the point of view of a nonclustered index, there are never any duplicate rows─even for identical data values─because nonclustered indexes include a unique row identification number along with the index value.) Only one of the IGNORE_DUP_ROW and ALLOW_DUP_ROW options can be set. The ALLOW_DUP_ROW option allows duplicate rows in a table. If ALLOW_DUP_ROW is set, you can create a new clustered index on a table that includes duplicate rows, and you can subsequently create duplicate rows with INSERT or UPDATE. If ALLOW_DUP_ROW is not set, the handling of duplicate rows depends on whether IGNORE_DUP_ROW is set. You cannot set ALLOW_DUP_ROW if IGNORE_DUP_ROW is set. If any index in the table is unique, the requirement for uniqueness─the most stringent requirement─takes precedence over the ALLOW_DUP_ROW option. Thus, ALLOW_DUP_ROW applies only to tables with nonunique indexes: you cannot use this keyword if a unique clustered index exists on any column in the table. (However, you might create a table with a nonunique clustered index and the ALLOW_DUP_ROW option, and then create a unique nonclustered index on the same table. If at that time the existing data does not include any duplicate rows, the CREATE UNIQUE NONCLUSTERED INDEX statement will succeed, but subsequent attempts to enter duplicate rows will fail because of the requirement for uniqueness on the new nonclustered index. If duplicate rows do exist when you attempt to create a unique nonclustered index, the statement fails.) The IGNORE_DUP_ROW option, like IGNORE_DUP_KEY, is used infrequently. Its most common function is to sort through a batch of data that includes errors you wish to eliminate. IGNORE_DUP_ROW controls what happens when you attempt to enter duplicate rows into a table in which they're not allowed. IGNORE_DUP_ROW applies only to tables with nonunique indexes: you cannot use this keyword if a unique index exists on any column in the table. If ALLOW_DUP_ROW is set, IGNORE_DUP_ROW is meaningless. If IGNORE_DUP_ROW is not set (and ALLOW_DUP_ROW is not set), the entire statement fails. (If there are other statements in the same transaction, they are executed normally.) If IGNORE_DUP_ROW is set, the row that would duplicate another row is not added or changed, and you get an error message. Changes to the database caused by the attempt to duplicate a row (for example, changes to index pages) are also backed out. Other rows are added or changed, or the rest of the index is created, as usual. ──────────────────────────────────────────────────────────────────────────── WARNING With IGNORE_DUP_ROW set, if you try to update a row in such a way that a duplicate row would be created, that row is discarded. Neither the new value nor the original value of the row that would have produced the duplicate exists in the updated table. For example, if you try to update "Smith" to "Jones" and "Jones" already exists, you will have one "Jones" and no "Smith." The disappearance of the row for "Smith" is distressing, but unavoidable, because an UPDATE statement is actually a DELETE followed by an INSERT. SQL Server has no way to know about the disallowed duplicate when it deletes the row, and the whole transaction can't be rolled back because the purpose of IGNORE_DUP_ROW (and of the IGNORE_DUP_KEY option) is to allow a transaction to proceed in spite of the presence of duplicates. ──────────────────────────────────────────────────────────────────────────── In the default situation, neither ALLOW_DUP_ROW nor IGNORE_DUP_ROW is set. Attempting to insert duplicate rows or create an index on a table with duplicate rows causes the entire statement to fail. The following table illustrates how ALLOW_DUP_ROW and IGNORE_DUP_ROW affect attempts to create a nonunique clustered index on a table that includes duplicate rows, and to enter duplicate rows into a table. Create index Insert duplicate rows (table has duplicate rows) ──────────────────────────────────────────────────────────────────────────── Neither Option Set Command fails. Command fails. ALLOW_DUP_ROW Set Command is completed. Command is completed. IGNORE_DUP_ROW Set Index created but All rows accepted except duplicate rows thrown duplicates; error message. out; error message. See the previous warning. ──────────────────────────────────────────────────────────────────────────── Comments Indexes speed data retrieval but can slow data update. Columns regularly used in joins should always be indexed. Creating a clustered index requires work-table space that is 1.21 times the size of the data. Create the clustered index before creating any nonclustered indexes, since nonclustered indexes are automatically rebuilt when a clustered index is created. Columns of bit, text, and image types cannot be indexed. You can create an index on a temporary table. It disappears when the table disappears. As the syntax signifies, you can create an index on a table in another database as long as you are the owner of that table. If there is no data in the table when an index is created, you should run UPDATE STATISTICS after data is added. A composite index, like any other index, is represented by one row in sysindexes. To display a report on an object's indexes, execute the sp_helpindex system procedure. Space is allocated to tables and indexes in increments of one extent, or 8 pages, at a time. In other words, an extent is allocated when the table or index is created, and another extent is allocated each time the previous extent becomes full. For a report on the amount of space allocated and used by an index, use sp_spaceused. Indexes cannot be created on columns containing text or image values. Permissions CREATE INDEX permission defaults to the table owner and is not transferable. See Also CREATE TABLE, DROP INDEX, INSERT, RECONFIGURE, SET, UPDATE, UPDATE STATISTICS, sp_helpindex, sp_spaceused CREATE PROCEDURE ──────────────────────────────────────────────────────────────────────────── Function Creates a stored procedure (a precompiled collection of SQL statements, often including control-of-flow language) that can take one or more user-supplied parameters. Syntax CREATE PROCedure [owner.]procedure_name [;number] [[(]@parameter_name datatype [= default] [, @parameter_name datatype [= default]]...[)]] [WITH RECOMPILE] AS SQL_statements Examples A. create procedure showind @tabname varchar(30) as select sysobjects.name, sysindexes.name, indid from sysindexes, sysobjects where sysobjects.name = @tabname and sysobjects.id = sysindexes.id Given a table name, the procedure showind (example A) displays its name and the names and identification numbers of any indexes on any of its columns. The following are the acceptable syntax forms for executing showind: execute showind titles execute showind @tabname = "titles" Or, if this is the first statement in a file or batch: showind titles B. create procedure showsysind @table varchar(30) = "sys%" as select sysobjects.name, sysindexes.name, indid from sysindexes, sysobjects where sysobjects.name like @table and sysobjects.id = sysindexes.id The procedure in example B displays information about the system tables if the user does not supply a parameter. C. create procedure showindnew @table varchar(30) = null as if @table is null print "Please give a table name" else select sysobjects.name, sysindexes.name, indid from sysindexes, sysobjects where sysobjects.name = @table and sysobjects.id = sysindexes.id The procedure in example C specifies an action to be taken if the parameter is NULL (that is, if the user does not give a parameter). Options procedure_name The name of the procedure. It must conform to the rules for identifiers. ;number An optional integer used to group procedures of the same name so that they can be dropped together with a single DROP PROCEDURE statement. Procedures used in the same application are often grouped this way. For example, the procedures used with the application orders might be named orderproc;1, orderproc;2, and so on. The statement DROP PROCEDURE orderproc would drop the entire group. Once procedures have been grouped, individual procedures within the group cannot be dropped. For example, the statement DROP PROCEDURE orderproc;2 is not allowed. parameter_name The name of a parameter to the procedure. The value of each parameter is supplied when the procedure is executed. (Parameter names are optional in CREATE PROCEDURE statements─a procedure need not take any parameters.) Parameter names must be preceded by the "at" symbol (@). The parameter name conforms to the rules for identifiers except that it can contain from 1 to 29 characters. Parameters are local to the procedure; the same parameter names can be used in other procedures. Parameters can only take the place of constants; they cannot be used in place of table names, column names, or the names of other database objects. However, the value of a parameter given at execution time can be an object name. If the value of a parameter is an object name and the object name is qualified by a database name or owner name, the entire name must be enclosed in single or double quotation marks. datatype The datatype of the parameter. A length in parentheses after the datatype entry is required for some datatypes. See the "Datatypes" section for a list of SQL-Server-supplied datatypes and their syntax. default A default parameter value for the procedure. If a default is defined, a user can execute the procedure without giving a parameter. The default must be a constant. It can include the wildcard characters (%, _, [ ], and [^]) if the procedure uses the parameter name with the LIKE keyword. (See example B.) The default can be NULL. The procedure definition can specify that some action be taken if the parameter value is NULL. (See example C.) WITH RECOMPILE Indicates that SQL Server will never save a plan for this procedure; the procedure will be recompiled each time it is executed. Use this optional clause when you expect that the parameters you supply to the procedure won't be typical, that is, they would not result in the same optimal plan. SQL_statements Statements that specify the actions the procedure is to take. Any number and kind of SQL statements can be included with the exception of CREATE statements. The SQL statements in a CREATE PROCEDURE statement can also include parameters. CREATE PROCEDURE statements often include control-of-flow language, including one or more of the following: DECLARE; IF...ELSE; WHILE; BREAK; CONTINUE; BEGIN...END; GOTO; RETURN; WAITFOR; /* comment */. The SQL statements can reference objects in another database as long as they are properly qualified. Comments Since stored procedures are compiled the first time they are executed, subsequent run time is much shorter than for the equivalent set of stand-alone statements. Most SQL Server database administration tasks are accomplished with predefined system procedures. Each of the system procedures' names begins with sp_. CREATE PROCEDURE statements cannot be combined with other SQL statements in a single batch. You can create a stored procedure in the current database only. A CREATE PROCEDURE statement cannot include any CREATE VIEW, CREATE DEFAULT, CREATE RULE, CREATE TRIGGER, or other CREATE PROCEDURE statements. The maximum number for each parameter, local variable, and global variable in a procedure is 40. (See "Variables (Local and Global)".) The maximum amount of text in a stored procedure is 65,280 characters. This limit is imposed by the fact that the text is stored in syscomments, where each procedure can occupy 255 rows of 256 bytes each. Once a procedure is created, you can run it by executing the EXECUTE statement along with the procedure's name and any parameters. Or, if a procedure is the first statement in a batch, you can give its name without the keyword EXECUTE. Any objects referenced in a procedure must already exist. You can create an object within a procedure and then reference it as long as the object is created before it is referenced. You can create a procedure that references a temporary object if the temporary object exists at the time the procedure is created or if the procedure itself creates the temporary object. A temporary object created in a procedure disappears when the procedure exits. If a procedure uses a temporary object that it does not create, you must re-create the temporary object and the stored procedure before you invoke that procedure. For this reason, system procedures (like sp_help) do not work on temporary objects. If you have a procedure that uses a temporary table and is not going to call other procedures, then it makes sense to use a temporary table. The table goes away when you exit the procedure. If you execute a procedure that calls another procedure, the called procedure can access objects created by the calling procedure. You can use the SET statement inside a stored procedure. The SET option you choose remains in effect during the execution of the procedure and then reverts to its former setting. Within a stored procedure, you cannot create an object (including a temporary one), drop it, and then create a new object with the same name. Inside a stored procedure, object names used with certain statements (the so-called utility programs) must be qualified with the object owner's name if other users are to use the stored procedure. For example, user mary, who owns table marytab, should qualify the name of her table inside a stored procedure (when it is used with these utility commands) if she wants other users to be able to execute it, like this: create procedure p1 as create index marytab_ind on mary.marytab(col1) This is because the object names are resolved when the procedure is run. In the example, if user john tried to execute procedure p1, SQL Server would look for a table called marytab owned by John. The utility statements are ALTER TABLE, CREATE TABLE, DROP TABLE, TRUNCATE TABLE, CREATE INDEX, DROP INDEX, UPDATE STATISTICS, and DBCC. Object names used with other statements (for example, SELECT or INSERT) inside a stored procedure need not be qualified, because the names are resolved when the procedure is compiled. The text of the CREATE PROCEDURE statement is stored in syscomments. To display it, execute the sp_helptext system procedure with the procedure name as the parameter. To display the text of a system procedure, execute sp_helptext from the master database. To rename a procedure, use sp_rename. For a report on the objects referenced by a procedure, use sp_depends. Permissions CREATE PROCEDURE permission defaults to the Database Owner, who can transfer it to other users. See Also Batch Queries, BEGIN...END, BREAK, Comments, CONTINUE, DECLARE, DROP PROCEDURE, EXECUTE, GOTO, GRANT, IF...ELSE, Parameters, RETURN, SELECT, System Procedures, Variables (Local and Global), WAITFOR, WHILE, Wildcard Characters, sp_depends, sp_helptext, sp_rename CREATE RULE ──────────────────────────────────────────────────────────────────────────── Function Specifies the domain of acceptable values for a particular column or for any column of a user-specified datatype. Syntax CREATE RULE [owner.]rule_name AS condition_expression Examples A. create rule limit as @advance < $1000 B. create rule pubid_rule as @pub_id in ('1389', '0736', '0877') C. create rule picture as @value like '_ _-%[0-9]' Options rule_name The name of the new rule. It must conform to the rules for identifiers. condition_expression The conditions that define the rule. It can be any expression that is valid in a WHERE clause and can include arithmetic operators, relational operators, IN, LIKE, BETWEEN, and so on. However, it cannot reference any column or other database object. Built-in functions that do not reference database objects can be included. A condition_expression takes one parameter. The parameter is prefixed by the "at" symbol (@) and refers to the value that is entered with the UPDATE or INSERT statement. You can use any name or symbol to represent the value when you write the rule, but the first character must be "@". Comments Rules do not apply to the data already existing in the database at the time the rules are created. CREATE RULE statements cannot be combined with other SQL statements in a single batch. You can create a rule in the current database only. After you've created a rule, use the sp_bindrule system procedure to link the rule to a column or user datatype. Rules cannot be bound to SQL-Server-supplied datatypes. The rule must be compatible with the datatype of the column. You cannot use "@value like A%", for example, as a rule for a numeric column. ──────────────────────────────────────────────────────────────────────────── NOTE If the rule is not compatible with the column to which you've bound it, SQL Server generates an error message when it tries to insert a value (not when you bind it). ──────────────────────────────────────────────────────────────────────────── Be sure to enclose character and date constants in quotation marks and to precede binary constants with "0x". Rule definitions and binding information are stored in system tables. The text of CREATE RULE statements is stored in the syscomments system table. To display it, execute the sp_helptext system procedure with the rule name as the parameter. After a rule is bound to a particular column or user datatype, its ID is stored in the syscolumns or systypes system tables. To get a report on a rule, use sp_help. To rename a rule, use sp_rename. You must drop a rule before you can create a new one of the same name, and you must unbind a rule (with the sp_unbindrule system procedure) before you drop it. You can bind a new rule to a column or datatype without unbinding the old one; the new rule overrides the old one. When unbinding a rule from a column, use sp_unbindrule. You can bind one rule to a user datatype and another rule to a column of that datatype. The rule most recently bound takes precedence. If a column has both a default and a rule associated with it, the default must fall within the domain defined by the rule. A default that conflicts with a rule will never be inserted. SQL Server generates an error message each time it attempts to insert the default. Permissions CREATE RULE permission defaults to the Database Owner, who can transfer it to other users. See Also Batch Queries, CREATE DEFAULT, DROP DEFAULT, DROP RULE, sp_bindrule, sp_help, sp_helptext, sp_rename, sp_unbindrule CREATE TABLE ──────────────────────────────────────────────────────────────────────────── Function Creates new tables. Syntax CREATE TABLE [[database.]owner.]table_name (column_name datatype [NOT NULL | NULL] [, column_name datatype [NOT NULL | NULL]]...) Examples create table titles (title_id tid, title varchar(80), type char(12), pub_id char(4), price money null, advance money null, royalty float null, ytd_sales money null, notes varchar(200), pubdate datetime) Options table_name The name of the new table. It must conform to the rules for identifiers and must be unique within this database and owner. You can create temporary tables by preceding the table name with a pound sign (#). If you wanted the table in example A to be a temporary table, you would call it #titles. Temporary tables are stored in tempdb..sysobjects by their names plus a system-supplied numeric suffix. Temporary tables disappear at the end of the current session or when they are explicitly dropped. You can create a table in a different database as long as you are listed in the sysusers table of the other database and have CREATE TABLE permission in that database. For example, to create a table called newtable in the database otherdb, use the following syntax: A. create table otherdb..newtable B. create table otherdb.yourname.newtable column_name The name of the column in the table. It must conform to the rules for identifiers and must be unique in the table. datatype The datatype of the column. System or user-defined datatypes are acceptable. Some datatypes require a length in the form datatype(n) where n is the length. To assign a datatype, use the format column_name datatype. (See "Datatypes" for more information.) NOT NULL Indicates that if there is no default for this column, an error message will be produced if no entry is made at insert time. Furthermore, the user cannot assign the value NULL in this column. If neither NULL nor NOT NULL is specified, NOT NULL is assumed. This SQL Server enhancement speeds retrievals and updates. NULL Indicates that if the user does not make an entry at insert time and there is no default entry for this column, SQL Server will assign the value NULL. Comments There can be up to 2 billion tables per database and 250 columns per table. The table is created in the currently open database unless a different database is explicitly specified in the CREATE TABLE statement with the optional database and owner names. (Cross-database creation of tables and indexes is allowed as long as the creator is listed in the sysusers table of the other database and has CREATE TABLE permission in that database. However, cross-database creation of views, rules, defaults, stored procedures, and triggers is not allowed.) User-defined datatypes are defined in terms of system datatypes. User datatypes permit frequently used type information to be accessed by a name you choose with a specified rule, default, and display format attached to it. User-defined datatypes are created with the sp_addtype system procedure before they are used in a table definition. You cannot create a row that has more than 1962 bytes of data. When you create user-defined datatypes with sp_addtype, there are three parameters: the name you assign the datatype; the base datatype (one of the SQL-Server-supplied datatypes─include length for char, varchar, binary, and varbinary); and either NULL or NOT NULL. The last parameter is optional. If you don't type NULL or NOT NULL, NOT NULL is assumed. The first two examples here are NOT NULL; the third is NULL. (Quotation marks are used when parameters contain blanks or punctuation marks.) A. sp_addtype tid1, "char(6)" B. sp_addtype tid2, "char(6)", "not null" C. sp_addtype tid3, "char(6)", null The NULL/NOT NULL assignment of a user datatype can be changed when you use it in a CREATE TABLE statement. However, the length specification cannot be changed: SQL Server will not allow you to specify a new length for a user-defined datatype in a CREATE TABLE statement. User-defined datatypes are removed with the sp_droptype system procedure. All of the datatypes that handle numbers (int, smallint, tinyint, money, and float) support arithmetic operations and aggregate functions (except that you cannot use modulo with float or money). Relational operators can be used on any of the datatypes except text and image. You can use the LIKE keyword and wildcard characters with char, varchar, text, and datetime data (but not to search for seconds or milliseconds). Built-in functions for concatenation and for finding substrings are provided for char, varchar, binary, and varbinary datatypes. (See "String Functions.") Conversion from one datatype to another is available with the CONVERT statement. (See "Conversion Function.") You can convert text data to varchar, and image data to varbinary, but it will be truncated to 255 bytes. When you create NULL columns with certain datatypes, SQL Server automatically converts them to a different internal datatype to allow the storage of null values. The char datatype is automatically converted to varchar; binary to varbinary; datetime to datetimn; float to floatn; int, smallint, and tinyint to intn; and money to moneyn. SQL Server does not inform the user of the type change, and the user need be concerned about it only if querying the system tables. Temporary tables are stored in the temporary database (tempdb), and permanent ones in the current database. You cannot create a view on a temporary table, nor can you associate rules, defaults, or triggers with temporary tables. You can use a user-defined datatype when creating a temporary table only if that type is in tempdb..systypes. (There are two ways to add a user-defined datatype (or any other object) to tempdb. To add an object for the current session only, execute sp_addtype while using the tempdb database. To add an object permanently, execute sp_addtype while using the model database and then restart SQL Server so that model is copied to tempdb.) Within a stored procedure, you cannot create a table, drop it, and then create a new table with the same name. For a report on a table and its columns, execute the sp_help system procedure. To rename a table, use sp_rename. ──────────────────────────────────────────────────────────────────────────── WARNING Procedures, triggers, and views that depend on an object whose name has been changed work properly until they are recompiled. However, recompilation takes place for many reasons and without notification to the user. When the procedure, trigger, or view is recompiled by SQL Server, it will no longer work. The user must change the procedure, trigger, or view text to reflect the new name of the dependent object. The safest course is to change the definitions of any dependent objects when you execute sp_rename. ──────────────────────────────────────────────────────────────────────────── Space is allocated to tables and indexes in increments of one extent, or 8 pages, at a time. In other words, an extent is allocated when the table or index is created, and another extent is allocated each time the previous extent becomes full. For a report on the amount of space allocated and used by a table, use sp_spaceused. For a report on the views and stored procedures that depend on a table, use sp_depends. You can define primary, foreign, and common keys on a table with the sp_primarykey, sp_foreignkey, and sp_commonkey system procedures. For a report on the keys that have been defined, use sp_helpkey. For a report on frequently used joins, use sp_helpjoins. Permissions CREATE TABLE permission defaults to the Database Owner, who can transfer it to other users. Any user can create temporary tables. See Also CREATE INDEX, CREATE RULE, CREATE VIEW, Datatypes, DROP INDEX, DROP RULE, DROP TABLE, sp_addtype, sp_commonkey, sp_depends, sp_foreignkey, sp_help, sp_helpjoins, sp_primarykey, sp_rename, sp_spaceused CREATE TRIGGER ──────────────────────────────────────────────────────────────────────────── Function Creates a trigger, which is a special kind of stored procedure often used for enforcing integrity constraints. A trigger is executed automatically when a user attempts the specified data modification statement on the specified table. Syntax CREATE TRIGGER [owner.]trigger_name ON [owner.]table_name FOR {INSERT | UPDATE | DELETE} [,{INSERT | UPDATE | DELETE}...] AS SQL_statements | IF UPDATE (column_name) [{AND | OR} UPDATE (column_name)...] Examples A. create trigger reminder on titles for insert, update as print "Don't forget to print a report for accounting." Example A prints a message every time anyone tries to add data or change data in the titles table. B. create trigger t1 on titleauthor for insert as if (select count(*) from titles, inserted where titles.title_id = inserted.title_id) = 0 begin print "Please put the book's title_id in the titles table first." rollback transaction end Example B prevents insertion of a new row into titleauthor if there is no corresponding title_id in the titles table. C. create trigger t2 on publishers for update as if update (pub_id) begin update titles set titles.pub_id = inserted.pub_id from titles, deleted, inserted where deleted.pub_id = titles.pub_id end If the pub_id column of the publishers table is changed, example C makes the corresponding change in the titles table. D. create trigger t3 on titleauthor for delete as begin delete titles from titles, deleted where deleted.title_id = titles.title_id delete titleauthor from titleauthor, deleted where deleted.title_id = titleauthor.title_id print "All references to this title have been deleted from titles and titleauthor." end If any row is deleted from titleauthor, that title is also deleted from the titles table (example D). If the book was written by more than one author, other references to it in titleauthor are also deleted. Options trigger_name The name of the trigger. It must conform to the rules for identifiers and must be unique in the database. INSERT, DELETE, UPDATE Keywords that can be included in any combination. SQL_statements Statements that specify trigger conditions and trigger actions. Trigger conditions specify additional criteria that determine whether the attempted INSERT, DELETE, or UPDATE will cause the trigger action(s) to be carried out. Trigger conditions often include a subquery preceded by the keyword IF. The trigger actions specified in the SQL statements go into effect when the user action (UPDATE, INSERT, or DELETE) is attempted. If multiple trigger actions are specified, they are grouped with BEGIN and END. Triggers can include any number and kind of SQL statements except SELECT statements; a trigger cannot return data to the user. The SQL statements in a trigger are similar to those in a stored procedure in that they often include control-of-flow language. A few special names are used in CREATE TRIGGER statements: ■ Deleted and inserted are logical (conceptual) tables. They are structurally like the table on which the trigger is defined─that is, the table on which the user action is attempted─and hold the old values or new values of the rows that would be changed by the user action. Deleted and inserted tables can be examined by the trigger to determine whether or how the trigger action(s) should be carried out. Deleted tables are used with DELETE and UPDATE; inserted with INSERT and UPDATE. See examples. ■ IF UPDATE is used to test if the specified column has been modified in any way. This allows specified trigger actions to be associated with updates to specified columns. See example B. More than one column can be specified. Use IF UPDATE like this: IF UPDATE (column_name) [{AND | OR} UPDATE (column_name)...] Comments Triggers are commonly used to enforce referential integrity (integrity rules about relationships between the primary and foreign keys of tables or views), to supply cascading deletes, and to supply cascading updates. (See examples B, C, and D, respectively.) You can define primary, foreign, and common keys on a table or view with the sp_primarykey, sp_foreignkey, and sp_commonkey system procedures. For a report on the keys that have been defined, use sp_helpkey. Trigger conditions specify additional criteria that determine whether the attempted INSERT, DELETE, or UPDATE will cause the trigger action(s) to be carried out. In example B (and in most triggers with trigger conditions), the subquery that follows the keyword IF is the trigger condition. Trigger actions specify the action(s) taken by SQL Server when the user action (that meets the trigger condition, if one is included) is attempted. Multiple statements are grouped with BEGIN and END. You can create a trigger in the current database only. If you use an owner name to qualify a trigger, you must explicitly qualify the table name the same way. A trigger may reference objects outside the current database. Triggers cannot be nested. That is, one trigger can't call another. If a trigger changes a table on which there is another trigger, the second trigger won't be executed. If you need a trigger to cause two actions, try to put both of the actions into a single trigger. (Or consider using a stored procedure instead.) For example, you might create an INSERT trigger on the sales table that updates titles.ytd_sales. You'd also like a trigger on updates to titles.ytd_sales that checks the royalty column in titles and changes it if necessary. The correct method is to write a single trigger that takes both actions─the update of titles.ytd_sales and the (possible) update of titles.royalty. This is because triggers cannot be nested─that is, one trigger cannot cause the execution of a second trigger. In performance terms, trigger overhead is usually very low. The time involved in running a trigger is spent mostly in referencing other tables, which may be either in memory or on the disk. The deleted and inserted tables, often referenced by triggers, are always in memory rather than on the disk because they are logical tables. The location of other tables referenced by the trigger determines the amount of time the operation takes. When an INSERT or UPDATE statement is executed, rows are added to the trigger table and to inserted at the same time. So the rows in inserted are always duplicates of one or more rows in the trigger table. Once a trigger is defined, the user action it specifies on the table to which it applies is always implicitly part of a transaction, along with the trigger itself. Triggers are often used to roll back that transaction if an error is detected. An UPDATE or INSERT trigger can make use of the IF UPDATE statement. It is used to test if the UPDATE or INSERT changed a particular column. IF UPDATE(column_name) is true for an INSERT statement whenever the column is assigned a value in the select list or in the VALUES clause. An explicit NULL or a default assigns a value to a column and thus activates the trigger. An implicit NULL, however, does not. Here are some examples: CREATE TABLE junk (a int null, b int not null) A. INSERT junk (a, b) VALUES (1, 2) IF UPDATE is true for either column B. INSERT junk VALUES (1, 2) IF UPDATE is true for either column C. INSERT junk VALUES (NULL, 2) Explicit NULL IF UPDATE is true for either column D. INSERT junk (b) VALUES (2) With a default for column a, IF UPDATE is true for either column E. INSERT junk (b) VALUES (2) With no default for column a, IF UPDATE is not true for column a IF UPDATE is never true for a DELETE statement. The text of the CREATE TRIGGER statement is stored in syscomments. To display it, execute the sp_helptext system procedure. The execution plan for the trigger is stored in sysprocedures. Each trigger is assigned an identification number, which is stored as a new row in sysobjects and as an entry in the sysobjects row for the table to which it applies. You can use the SET statement inside a trigger. The SET option you invoke remains in effect during the execution of the trigger and then reverts to its former setting. It is recommended that a trigger not include SELECT statements that return results to the user, since special handling for these returned results would have to be written into every application program in which modifications to the trigger table are allowed. You cannot create a trigger on a view or on a temporary object. A TRUNCATE TABLE statement is not caught by a DELETE trigger. Although a TRUNCATE TABLE statement is, in effect, like a DELETE without a WHERE clause (it removes all rows), it is not logged, and so cannot execute a trigger. Since permission for the TRUNCATE TABLE statement defaults to the table owner and is not transferable, only the table owner should be concerned about inadvertently circumventing a DELETE trigger with a TRUNCATE TABLE statement. When you drop a table, any triggers associated with it are also dropped. A trigger cannot apply to more than one table. However, the same trigger action can be defined for all three user actions in the same CREATE TRIGGER statement. A table can have a maximum of three different triggers─one each for INSERT, UPDATE, and DELETE. For a report on a trigger, execute the sp_help system procedure. For a report on the tables and views that are referenced by a trigger, use sp_depends. You can rename a trigger with sp_rename. Permissions CREATE TRIGGER permission defaults to the table owner and is not transferable. See Also CREATE PROCEDURE, DROP TRIGGER, sp_commonkey, sp_depends, sp_foreignkey, sp_help, sp_helptext, sp_primarykey, sp_rename, sp_spaceused CREATE VIEW ──────────────────────────────────────────────────────────────────────────── Function Creates views (alternative ways of looking at the data in one or more tables). Syntax CREATE VIEW [owner.]view_name [(column_name [, column_name]...)] AS select_statement Examples A. create view titles_view as select title, type, price, pubdate from titles B. create view accounts (title, advance, amt_due) as select title, advance, price * royalty * ytd_sales from titles where price > $5 C. create view cities (authorname, acity, publishername, pcity) as select au_lname, authors.city, pub_name, publishers.city from authors, publishers where authors.city = publishers.city D. create view cities2 as select authorname = au_lname, acity = authors.city, publishername = pub_name, pcity = publishers.city from authors, publishers where authors.city = publishers.city Options view_name The name of the new view. It must be a legal identifier. column_name The names to be used for the columns in the view. This is always legal but only necessary when a column is derived from an arithmetic expression, a function, or a constant; when two or more columns could otherwise have the same name (usually because of a join); or when you want to give a column in a view a different name than the column from which it is derived. (See example C.) Column names can also be assigned in the SELECT statement. (See example D.) If no column names are specified, the view columns acquire the same names as the columns in the SELECT statement. select_statement The SELECT statement that defines the view. It can use more than one table and other views. Comments You cannot include ORDER BY or COMPUTE clauses or the DISTINCT or INTO keywords in the SELECT statements that define views. You cannot create a trigger on a view. You can create a view in the current database only. Currently there are no temporary views. If you alter the structure of a view's underlying table(s) by adding columns, the new columns will not appear in a view defined with a "SELECT *" clause unless the view is deleted and redefined. The asterisk shorthand is interpreted and expanded when the view is first created. If a view depends on a table (or view) that has been dropped, SQL Server produces an error message if anyone tries to use the view. If a new table (or view) is created to replace the one that has been dropped, the view again becomes usable. You cannot create a view on a temporary table. You can redefine a view without redefining other views that depend on it unless the redefinition makes it impossible for SQL Server to translate the dependent view. CREATE VIEW statements cannot be combined with other SQL statements in a single batch. There are no restrictions on querying through views, but there are some restrictions on data update through views. Data update statements may not change any column in a view that is a computation. Data update statements may not change a view that includes aggregate functions (built-in functions and a GROUP BY or COMPUTE BY clause). INSERT statements are not allowed unless all NOT NULL columns in the underlying table or view are included in the view through which you are inserting new rows. (SQL Server has no way to supply values for NOT NULL columns in the underlying table or view.) INSERT and UPDATE statements are not allowed unless the columns being updated all belong to the same base table. (It's illegal in SQL to use data update statements on more than one table in a single statement.) However, if a view contains columns from more than one table but the data update statement references columns from only one of those tables, the statement is legal. When you query through a view, SQL Server checks to make sure that all the database objects referenced anywhere in the statement exist, that they are valid in the context of the statement, and that data update statements do not violate data integrity rules. If any of these checks fail, you get an error message. If the checks are successful, it translates the view into an action on the underlying table(s). You can use views as security mechanisms by granting permission on a view but not on underlying tables. The text of the CREATE VIEW statement is stored in syscomments. To display it, execute the sp_helptext system procedure with the view name as the parameter. You can rename a view with sp_rename. To get a report of the tables or views on which a view depends, and of objects that depend on a view, execute the sp_depends system procedure. See SQL Server Learning TRANSACT-SQL for more information and examples on views. Permissions CREATE VIEW permission defaults to the Database Owner, who can transfer it to other users. The creator of a view must have SELECT permission on any objects referenced in the view definition. See Also DROP VIEW, Views, sp_depends, sp_help, sp_helptext, sp_rename Datatypes ──────────────────────────────────────────────────────────────────────────── Function Specify data characteristics of columns, stored procedure parameters, and local variables. For text and image datatypes used to store long string of data, see "Text/Image Datatypes." Syntax column_name text column_name image column_name timestamp {param_name | variable_name | column_name} int {param_name | variable_name | column_name} smallint {param_name | variable_name | column_name} tinyint {param_name | variable_name | column_name} float {param_name | variable_name | column_name} char(n) {param_name | variable_name | column_name} varchar(n) {param_name | variable_name | column_name} binary(n) {param_name | variable_name | column_name} varbinary(n) {param_name | variable_name | column_name} bit {param_name | variable_name | column_name} money {param_name | variable_name | column_name} datetime {param_name | variable_name | column_name} sysname {param_name | variable_name | column_name} user_type_name Datatypes include system datatypes and user-defined datatypes. SQL Server supplies the sysname user datatype. The pubs sample database also supplies user datatypes. You can also create your own user datatypes. Some datatypes require a length. Examples A. create table food (fname varchar(20), fcals int, frate int) B. create procedure auname_sp @auname varchar(40) as select au_lname, title, au_ord from authors, titles, titleauthor where @auname = au_lname and authors.au_id = titleauthor.au_id and titles.title_id = titleauthor.title_id C. declare @hope money Options param_name The name of a stored procedure parameter. It must conform to the rules for identifiers. variable_name The name of a local variable. It must conform to the rules for identifiers. column_name The name of a database column. It must conform to the rules for identifiers and must be unique within its table. int An integer column that holds whole numbers between 231- 1 (that is, 2,147,483,647) and -231+ 1 (that is, -2,147,483,648), inclusive. However, you cannot enter -2,147,483,648, but you can enter -2,147,483,647 - 1. You can store this number or it can be the result of a calculation. Storage size is 4 bytes. smallint A small integer column that holds whole numbers between 215- 1 (32,767) and -215 (-32,768), inclusive. Storage size is 2 bytes. tinyint A tiny integer column that holds whole numbers between 0 and 255, inclusive. Storage size is 1 byte. float A floating-point column that holds positive or negative floating-point numbers. This column has 15-digit precision. The range of values is approximately 1.7E - 308 to 1.7E + 308. Storage size is 8 bytes. All the arithmetic operations except modulo are available with float. Data of type float can include an optional exponent. For float data with an exponential component, enter a number (with or without a decimal point and a positive or negative sign), followed by "e" or "E", followed by an optionally signed integer. The value represented by a float is the product of the first number and the power of 10 of the second number. char(n) A character column that any combination of up to 255 letters, symbols, and numbers. Specify the maximum size of the column with n. A char column can contain 0 characters, but n must be between 1 and 255. Storage size is n, no matter what the actual entry length. Columns of type char are accessed a little faster than varchar columns and use more storage space. Choose char when you know the size of character data (as in a social security number or a zip code). Data of type char must be enclosed in single or double quotation marks when it's input. Numbers stored in char columns must be converted to a numeric data type before you can do arithmetic on them. You can use the LIKE keyword and wildcard characters with char, varchar, and datetime data (but not to search for seconds or milliseconds). If you enter strings that are too long for the specified length, char and varchar entries are truncated. varchar(n) A column of variable-length characters that holds any combination of up to 255 letters, symbols, and numbers. Specify the maximum size of the column with n. A varchar column can contain 0 characters, but n must be between 1 and 255. Storage size is the actual size of the data values entered, not n. Because of the way it uses space, varchar is best suited for data-like names where the length of each entry can vary considerably. Data of type varchar must be enclosed in single or double quotation marks when it's input. You can use the LIKE keyword and wildcard characters with char, varchar, and datetime data (but not to search for seconds or milliseconds). If you enter strings that are too long for the specified length, char and varchar entries are truncated. text Variable-length columns that can hold up to 2,147,483,647 (231- 1) characters. The text datatype may not be used for variables or parameters in stored procedures. (See "Text/Image Datatypes.") binary(n) A column that holds up to 255 bytes of fixed-length binary data. The binary datatypes are not for hexadecimal data, but rather for bit patterns. They are used for storing programming code or pictures, not for hexadecimal numbers. Specify the maximum byte length of the column with n. A binary column can contain 0 bytes, but n must be between 1 and 255. Storage size is n no matter what the actual entry length. The maximum binary value is 255 bytes, each of which is "ff". Columns of type binary are accessed a little faster than varbinary columns and use more storage space. Choose binary when you think the data entries in the column will be close to the same size. Data of type binary consists of the characters 0 to 9 and A to F (or a to f), and the first binary string must be preceded with "0x" when it's input. For example, to input "FF", type 0xFF. A length of 10 for a binary or varbinary column means 20 characters. If you enter strings that are too long for the specified length, binary and varbinary entries are truncated. varbinary(n) A variable-length binary column that holds up to 255 bytes of variable-length binary data. The binary datatypes are not for hexadecimal data, but rather for bit patterns. They are used for storing programming code or pictures, not for hexadecimal numbers. Specify the maximum size of the column with n. A varbinary column can contain 0 bytes, but n must be between 1 and 255. Storage size is the actual size of the data values entered, not n. Because of the way it uses space, choose varbinary when you expect variation in the size of the data. Data of type varbinary consists of the characters 0 to 9 and A to F (or a to f), and the first varbinary string must be preceded with "0x" when it's input. A length of 10 for a binary or varbinary column means 20 characters. If you enter strings that are too long for the specified length, binary and varbinary entries are truncated. image A variable-length column that holds between 0 and 2,147,483,647 bytes of binary data. The image datatype may not be used for variables or parameters in stored procedures. (See "Text/Image Datatypes" for details.) bit A column that holds either a 1 or a 0. Integer values other than 1 or 0 are accepted but are always interpreted as 1. Storage size is 1 bit. Multiple bit types in a table can be collected into bytes. For example, 7 bit columns fit into 1 byte; 9 bit columns take 2 bytes. The status column in the syscolumns system table indicates the unique offset position for bit columns. Columns of type bit cannot be NULL and cannot have indexes on them. Use bit for true/false or yes/no types of data. money A column that stores dollar and cent values between ±922,337,203,685,477.5807 dollars with accuracy to a ten-thousandth of a dollar. Money values are represented as double-precision integers. Storage size is 8 bytes. Use money for all decimal currency entries. Data of type money should be preceded with a dollar sign ($) when it's input. If there is no dollar sign, SQL Server treats the value as a float. Depending on the exact input, SQL Server may not accept it, or may lose some of the value's precision. To represent negative money values, a minus sign (-) must be placed after the dollar sign. You cannot enter money values with commas, although the default print format for money data places a comma after every three digits. When money values are displayed, they are rounded up to the nearest cent. All the arithmetic operations except modulo are available with money. datetime A column that holds dates and times of day. Storage size is two 4-byte integers: 4 bytes for the number of days before or after the base date of January 1, 1900, and 4 bytes for the number of milliseconds after midnight. The date segment of datetime values representing dates before the base date are stored as negative values. Data values for datetime range from January 1, 1753, to December 31, 9999. The maximum datetime value is the last day of the year 9999. SQL Server will reject all values it cannot recognize as dates between 1753 and 9999. SQL Server stores datetime data to an accuracy of .003 milliseconds. The entries .001, .002, and .003 are stored as .000, while an entry of .004 is stored as .003. The default display format for dates is "mon dd yyyy hh:miAM" (or PM), such as "Apr 15 1988 10:23AM". When you enter datetime values, always enclose them in single or double quotation marks. Case is always ignored, and spaces can occur anywhere between date parts. SQL Server recognizes a wide variety of data entry formats for date and time values. When time data is entered, the order of time components is significant: hours; then minutes; then seconds; then milliseconds; then AM, am, PM, or pm (12AM is midnight, 12PM is noon). To be recognized as time, a value must contain either a colon or an AM/PM signifier. Milliseconds can be preceded either with a colon or a period. If preceded by a colon, the number means thousandths of a second. If preceded by a period, a single digit means tenths of a second, two digits mean hundredths of a second, and three digits mean thousandths of a second. For example, "12:30:20:1" means twenty and one-thousandth second past 12:30; "12:30:20.1" means twenty and one-tenth second past 12:30. The following formats for time data are acceptable: 14:30 14:30[:20:999] 14:30[:20.9] 4am 4 PM [0]4[:30:20:500]AM When date data is entered, the order of components can vary. Months can be in alphabetic or numeric format. Alphabetic format is either the full name of the month or its first three letters. Case is not significant. If the month is given in numeric format, you can use any of the following separators: slash (/), hyphen (-), period (.), or space ( ). You can also give a date without separators: you can enter January 1, 1998 as "19980101" or "980101". If the month is given in alphabetic format, the day and year can be in any location, either before or after the month. In addition to the previously mentioned separators, the day can optionally be followed by a comma (,). When the month is alphabetic and the values for day and year are grouped together, the first value is assumed to be the day. For example, 2 3 mar is stored as "Mar 2 2003 12:00:00AM". If the year is given with two digits, <50 is next century and >=50 is this century. So "25" is "2025" while "50" is "1950". The following formats for date data are acceptable: Apr[il] 15[,] [19]88 [15] Apr[il][,] [19]88 1988 APR[IL] 4/15/1988 4-15-88 [19]88[0415] A datetime value can be either time then date or date then time, or the year can be separated from the rest of the date: 10:23PM Apr 15 1988 Apr 15 1988 10:23PM Apr 15 10:23PM 1988 If time is missing, it defaults to midnight. For example, if you insert "Apr 15 1988", SQL Server stores this as "Apr 15 1988 12:00AM". If the day of the month is missing, it defaults to the first day of the month. If the month is missing, it defaults to January. If the entire date is missing, it defaults to the base date, January 1, 1900. For example, if you insert "4:33", SQL Server stores it as "Jan 1 1900 4:33AM". Use the CONVERT function's style parameter to obtain many additional date styles. (See "Conversion Function.") You can do some arithmetic calculations on datetime values with the built-in date functions. (See "Date Functions.") When you use LIKE with datetime values, SQL Server converts the dates to the standard datetime format and then to varchar. Since the standard storage format doesn't include seconds or milliseconds, you cannot search for seconds or milliseconds with LIKE and a match pattern. It is a good idea to use LIKE when you search for datetime values, since datetime entries may contain a variety of date parts. For example, if you insert the value "9:20" into a column named arrival_time, the clause WHERE arrival_time = "9:20 " would not find it because SQL Server converts the entry into "Jan 1, 1900 9:20AM". However, the clause WHERE arrival_time LIKE "%9:20%" would find it. sysname A user-defined datatype supplied by SQL Server and used in the system tables. Its definition is varchar(30) "NOT NULL". timestamp A column that is automatically updated when you use browse mode in a DB-LIBRARY application. (See "Browse Mode.") Every time a row containing a timestamp column is inserted or updated, the timestamp column is automatically updated. A table can have only one column of timestamp type. Its definition is varbinary (8) null. When a table is created that contains timestamp data, both the column and the datatype must be named "timestamp". user_type_name The name of a user-defined datatype created with the sp_addtype system procedure. For example, to add a type called nm, you might execute the following system procedure: sp_addtype nm, "varchar(30)", null The name of the user-defined datatype must conform to the rules for identifiers. Comments All system datatypes and user-defined datatypes default to NOT NULL in CREATE TABLE statements. When you include a user-defined datatype in a CREATE TABLE statement, you can override its nulltype specification. However, you cannot override the length specification. Datatypes are database objects, and their names must conform to the rules for identifiers. The names of the SQL-Server-supplied datatypes are not reserved words and can be used to name other objects. User-defined datatypes are defined in terms of system datatypes. Creating user-defined datatypes permits frequently used datatype and length information to be accessed under a name you choose. User-defined datatypes are created with the sp_addtype system procedure and dropped with sp_droptype. Rules and defaults can be bound to user datatypes with the sp_bindrule and sp_bindefault system procedures. See Chapter 2, "System Procedures," for details. When you create user-defined datatypes with sp_addtype, there are three parameters: the name you assign the datatype; the base datatype (one of the SQL-Server-supplied datatypes, which include length for char, varchar, binary, and varbinary); and either NULL or NOT NULL. The last parameter is optional. ──────────────────────────────────────────────────────────────────────────── WARNING Whenever you create a table containing timestamp data, both the column and the datatype must be named literally "timestamp". For example, the following table has one timestamp column create table dates (timestamp timestamp) Do not change the name of a timestamp column using the sp_addtype system procedure. If you do not use "timestamp" as the column name and datatype name, the timestamp is not set when rows are inserted into a table. ──────────────────────────────────────────────────────────────────────────── If you don't specify NULL or NOT NULL, NOT NULL is assumed. In the following examples, the first two examples are NOT NULL; the third is NULL. (Quotation marks are used when a parameter contains blanks or punctuation marks.) sp_addtype tid1, "char(6)" sp_addtype tid2, "char(6)", "not null" sp_addtype tid3, "char(6)", null The NULL/NOT NULL assignment for a user-defined datatype can be changed in a CREATE TABLE statement. However, the length specification cannot be changed; SQL Server will not allow you to specify a length for a user-defined datatype in a CREATE TABLE statement. User-defined datatypes cannot be used in temporary tables unless the datatypes exist in the tempdb database, that is, unless they have been explicitly created in tempdb during the current session or unless they have been permanently added to the model database by changing the instmodl.sql script. See the SQL Server Installation Guide for the location of this script on your system. All of the datatypes that handle numbers─int, smallint, tinyint, money, and float─support arithmetic operations and aggregate functions (except that you cannot use modulo with float or money). Computations that result in overflow conditions are handled as errors. Relational operators can be used on any of the datatypes except text and image. There are two ways to specify literal quotation marks within a char or varchar entry. The first method is to use two quotation marks. For example, if you have begun a character entry with a single quotation mark and wish to include a single quotation mark as part of the entry, use two single quotation marks: 'I don"t understand.' With double quotation marks: "He said, ""It's not really confusing.""" The second method is to enclose one quotation mark in the opposite kind of quotation mark. In other words, surround an entry containing double quotation marks with single quotation marks (or vice versa). Here are some examples: 'George said, "There must be a better way."' "Isn't there a better way?" 'George asked, "Isn"t there a better way?"' To continue a character string onto the next line of your screen, enter a backslash (\) before going to the next line. Built-in functions are provided for character, numeric, and date datatypes. (See "Date Functions," "Mathematical Functions," "String Functions," and "Text Functions" for details.) Many conversions from one datatype to another are handled automatically by SQL Server. Others are available with the CONVERT function. However, there are some type conversions that can never be done. For example, you cannot convert datetime to tinyint. (See "Conversion Function" for more information.) Text and image data can be converted to char or varchar and binary or varbinary, respectively. Use the style parameter of the CONVERT function to obtain a wide variety of date display formats. (See "Conversion Function" for more information.) When you create NULL columns with certain datatypes, SQL Server automatically converts them to a different internal datatype to allow the storage of null values. The char datatype is automatically converted to varchar; binary to varbinary; datetime to datetimn; float to floatn; int, smallint, and tinyint to intn; and money to moneyn. SQL Server does not inform the user of the type change, and the user should be concerned about it only if querying the system tables. For a report on a user-defined datatype, execute sp_help. You can rename a user-defined datatype with sp_rename. When you perform mixed-mode arithmetic─arithmetic operations on values of different datatypes, for example int + smallint ─the resultant datatype is determined by the hierarchy of datatypes. The hierarchy of datatypes is created from the rank ordering of the hexadecimal values of the datatype codes. In a mixed-mode expression, the lower type is converted internally to the higher type. In the following example, the qty from the sales table is multiplied by royalty from the roysched table. The qty is a smallint (INT2), and the royalty is an int (INT4). INT2 has a hexadecimal value of 34; INT4 has a hexadecimal value of 38. Therefore, the resultant datatype is an int (INT4). smallint(qty) * int(royalty) = int The hierarchy of datatypes is shown in the systypes system table. (Also see the SQL Server System Administrator's Guide and the "Conversion Function.") You can use the following query to obtain a list of datatypes with their decimal and hexadecimal values. Select the name of the datatype and its decimal type code, then use CONVERT to compute the hexadecimal value of each type. select name, type, hex=convert(varbinary(2),type) from systypes order by type <execute> ╓┌─┌───────────────────┌────────┌────────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── name type hex ------------ ------- ------- text 35 0x23 timestamp 37 0x25 varbinary 37 0x25 intn 38 0x26 sysname 39 0x27 ──────────────────────────────────────────────────────────────────────────── sysname 39 0x27 varchar 39 0x27 binary 45 0x2d image 46 0x2e char 47 0x2f tinyint 48 0x30 bit 50 0x32 smallint 52 0x34 int 56 0x38 money 60 0x3c datetime 61 0x3d float 62 0x3e floatn 109 0x6d moneyn 110 0x6e datetimn 111 0x6f (19 rows affected) The results of this query on one of your databases will be slightly different if you have installed any user-defined datatypes. User types take on the hexadecimal value of their base types. See Also Aggregate Functions, Conversion Function, CREATE PROCEDURE, CREATE TABLE, Date Functions, EXECUTE, Parameters, Text/Image Datatypes, Text/Image Functions, Variables (Local and Global), Wildcard Characters, sp_addtype, sp_bindefault, sp_bindrule, sp_droptype, sp_help, sp_rename, sp_unbindefault, sp_unbindrule Date Functions ──────────────────────────────────────────────────────────────────────────── Function Manipulate datetime values. Syntax All date functions take parameters except GETDATE. Function names, parameters, and results are as follows: ╓┌─────────┌────────────────────────────────┌────────────────────────────────╖ Function Parameter Result ──────────────────────────────────────────────────────────────────────────── DATEADD (datepart, number, date) A date produced by adding an interval to a date you specify. The result is a datetime value equal to the date plus the number of date parts. DATEDIFF (datepart, date1, date2) The number of dateparts between two specified dates. The result is a signed integer value equal to date2 minus date1 in date parts. Function Parameter Result ──────────────────────────────────────────────────────────────────────────── DATENAME (datepart, date) The specified datepart (the first parameter) of the specified date (the second parameter) as a character string. DATEPART (datepart, date) The specified datepart (the first parameter) of the specified date (the second parameter) as an integer. GETDATE ( ) The current date and time in SQL Server's standard internal format for datetime values. GETDATE takes the NULL parameter, ( ). ──────────────────────────────────────────────────────────────────────────── Function Parameter Result ──────────────────────────────────────────────────────────────────────────── ──────────────────────────────────────────────────────────────────────────── Examples A. Statement Result select rightnow = getdate( ) Nov 25 1995 10:32AM select datepart(month, getdate( )) 11 select datename(month, getdate( )) November Example A assumes a current date of November 25, 1995, 10:32 am. B. select interval = dateadd(day, 21, pubdate) from titles Example B displays the new publication dates when the publication dates of all the books in the titles table slip 21 days. C. select newdate = datediff(day, pubdate, getdate()) from titles Example C finds the number of days elapsed between pubdate and the current date (found with the GETDATE function). Options date A parameter used with DATEADD, DATEDIFF, DATENAME, and DATEPART. The date option can be either the function GETDATE, a character string in one of the acceptable date formats (see "Datatypes"), or the name of a datetime column. datepart A parameter used with DATEADD, DATEDIFF, DATENAME, and DATEPART. The following table lists the dateparts, the abbreviations recognized by SQL Server, and the acceptable values: ╓┌────────────┌─────────────┌────────────────────────────────────────────────╖ Date Part Abbreviation Values ──────────────────────────────────────────────────────────────────────────── year yy 1753-9999 quarter qq 1-4 month mm 1-12 day of year dy 1-366 day dd 1-31 week wk 1-53 weekday dw 1-7 (Sun.-Sat.) hour hh 0-23 minute mi 0-59 second ss 0-59 millisecond ms 0-999 ──────────────────────────────────────────────────────────────────────────── If the year is given with two digits, <50 is the next century and >=50 is this century. So "25" is "2025", while "50" is "1950". When you use week with DATEDIFF, you get the number of Sundays between the two dates, including the first date but not the second. For example, the difference in weeks between Sunday, January 4, and Sunday, January 11, is one. Milliseconds can be preceded either with a colon or a period. If preceded by a colon, the number means thousandths of a second. If preceded by a period, a single digit means tenths of a second, two digits mean hundredths of a second, and three digits mean thousandths of a second. For example, "12:30:20:1" means twenty and one-thousandth second past 12:30; "12:30:20.1" means twenty and one-tenth second past 12:30. Comments Date functions can be used in the select list or WHERE clause of a query. Use the datetime datatype only for dates after January 1, 1753. When you enter datetime values, always enclose them in single or double quotation marks. Values of type datetime are stored in a special internal format. They are displayed in a default format and can be entered in a variety of formats. (See "Datatypes" for more information.) SQL Server automatically converts between character and datetime values when necessary (for example, when you compare a character value to a datetime value). Use the style parameter of the CONVERT function to obtain a wide variety of date display formats. (See "Conversion Function" for more information.) The DATEDIFF function does not produce reliable results if your result in milliseconds is greater than 2 billion. For example, if you try to find the difference in milliseconds between two dates more than a year apart, your result will not be accurate. See Also Conversion Function, Expressions, Functions, SELECT, WHERE Clause DBCC ──────────────────────────────────────────────────────────────────────────── Function Checks the logical and physical consistency of a database, ordinarily either because damage is suspected or as part of a periodic check. Syntax DBCC {CHECKTABLE (table_name) | CHECKDB [(database_name)] | CHECKALLOC [(database_name)] | CHECKCATALOG [(database_name)] | DBREPAIR (database_name, DROPDB) } Examples A. dbcc checktable(titleauthor) B. dbcc checkalloc(pubs) C. dbcc dbrepair(pubs, dropdb) Options CHECKTABLE Checks the specified table to see that index and data pages are correctly linked, that indexes are in properly sorted order, that all pointers are consistent, that the data information on each page is reasonable, and that page offsets are reasonable. CHECKDB Runs the same checks as CHECKTABLE, but on each table in the specified database. If no database name is given, CHECKDB checks the current database. CHECKALLOC Checks the specified database to see that all pages are correctly allocated and that no page is allocated that isn't used. If no database name is given, CHECKALLOC checks the current database. CHECKALLOC reports on the amount of space allocated and used. For example, here's the report on pubs: Checking pubs Alloc page 0 (# of extent=32 used pages=65 ref pages=65) Alloc page 256 (# of extent=32 used pages=140 ref pages=140) Alloc page 512 (# of extent=32 used pages=186 ref pages=186) Alloc page 768 (# of extent=10 used pages=40 ref pages=40) Total (# of extent=106 used pages=431 ref pages=431) in this database DBCC execution completed. If DBCC printed error messages, see your system administrator. This database contains four allocation units, 512K each, or 2 megabytes in all (the default size of a database). There are 32 extents (required for the creation of new database objects) per allocation page; the number of extents tells whether or not you can create new objects in the database. There are 8 pages per extent; the number of used pages tells whether you can add data to existing objects. To get an approximation of the space used, multiply the total number of used pages by 2K. CHECKCATALOG Checks for consistency in and between system tables. For example, it makes sure that every type in syscolumns has a matching entry in systypes, that every table and view in sysobjects has at least one column in syscolumns, and that the last checkpoint in syslogs is valid. CHECKCATALOG also reports on any segments that have been defined. If no database name is given, CHECKCATALOG checks the current database. DBREPAIR (database_name, DROPDB) Drops a damaged database. The DROP DATABASE statement does not work on a damaged database. No user, including the person executing the statement, can be using the database that is being dropped when this DBCC statement is executed. Comments DBCC (Database Consistency Checker) can be run while the database is active except for the DBREPAIR(database_name, DROPDB) option. A warning message may be issued by DBCC, either during CHECKDB or CHECKTABLE, stating that the value in the dpages column of the sysindexes row is not the same as DBCC's count of data pages for some table. To correct this inconsistency, follow this procedure: 1. Use the sp_configure stored procedure to enable updates to system catalogs. 2. Execute the RECONFIGURE WITH OVERRIDE statement for that change in configuration to take effect. 3. When using the database where the DBCC check was done, execute the following statement: update sysindexes set dpages=<count from DBCC> where id=<id of table> and indid=<0 or 1> 4. Execute the CHECKPOINT statement in the database where the DBCC check was done. 5. Disable updates to system catalogs with sp_configure. 6. Execute the RECONFIGURE statement. Because of the special status of sysindexes as a system table, it is necessary to qualify the update and restrict it to a single row as described above. The value for the qualification's comparison to indid will be 1 if there is a clustered index on that table and 0 if there is none. Permissions DBCC permission defaults to the table owner for CHECKTABLE and to the Database Owner for CHECKDB, CHECKALLOC, CHECKCATALOG, and DBREPAIR. It is not transferable. See Also DROP DATABASE, sp_helpdb DECLARE ──────────────────────────────────────────────────────────────────────────── Function Declares the name and type of local variables for a batch or procedure. Local variables are assigned values with a SELECT statement. Syntax Variable declaration: DECLARE @variable_name datatype [, @variable_name datatype...] Variable assignment: SELECT @variable = expression [, @variable = expression...] [FROM table_list] [WHERE search_conditions] [GROUP BY group_by_list] [HAVING search_conditions] [ORDER BY order_by_list] [COMPUTE function_list [BY by_list]] Examples A. declare @one varchar(18), @two varchar(18) select @one = "this is one", @two = "this is two" if @one = "this is one" print "you got one" if @two = "this is two" print "you got two" else print "nope" B. declare @veryhigh money select @veryhigh = max(price) from titles if @veryhigh > $20 print "Ouch!" Options variable_name The name of the variable. It must conform to the rules for identifiers. Variables can be used only in place of constants. They cannot be used in place of the names of tables, columns, other database objects, or keywords. datatype A system datatype or a user datatype. Comments The "at" sign (@) denotes a variable name. Local variables must be used within the batch or procedure in which they are declared. Local variables are often used in a batch or procedure as counters for WHILE loops or IF...ELSE blocks. When they are used in stored procedures, local variables are declared for automatic, noninteractive use by the procedure when it executes. The SELECT statement that assigns a value to the local variable usually returns a single value. If the SELECT variable assignment statement returns more than one value, the variable is assigned the last value returned. The PRINT and RAISERROR statements can take local variables as parameters. The SELECT statement that assigns values to variables cannot be used to retrieve data in the same statement. The maximum number for each parameter and global variable in a procedure is 255. See "Variables (Local and Global)." There is no limit on the number of local variables. Users cannot declare global variables. See Also Parameters, PRINT, RAISERROR, Variables (Local and Global), WHILE DELETE ──────────────────────────────────────────────────────────────────────────── Function Removes rows from a table. Syntax A. DELETE [[database.]owner.]{table_name | view_name} [FROM [[database.]owner.]{table_name | view_name} [, [[database.]owner.]{table_name | view_name}]...] [WHERE search_conditions] B. DELETE [FROM] [[database.]owner.]{table_name | view_name} [WHERE search_conditions] Examples A. delete authors Example A deletes all rows from the authors table. B. delete from authors where au_lname = "McBadden" Example B deletes one row from the authors table. C. delete titles from titles, authors, titleauthor where authors.au_lname = 'Bennet' and authors.au_id = titleauthor.au_id and titleauthor.title_id = titles.title_id Example C deletes rows for books written by Bennet from the titles table. (The pubs database includes a trigger (deltitle) that prevents the deletion of titles that are recorded in the sales table; you must drop this trigger to make the last example work.) Options FROM (see syntax A) The additional FROM optionally lets you name more than one table or view to use with a WHERE clause to specify which rows to delete. The additional FROM clause allows you to delete rows from one table based on data stored in other tables, giving you much of the power of an embedded SELECT statement. FROM (see syntax B) An optional keyword for compatibility with other versions of SQL. Follow it with the name of the table or view from which you want to remove rows. WHERE A standard WHERE clause. (See "WHERE Clause" for more information.) Comments If you don't use a WHERE clause, all rows in the table named after DELETE [FROM] are removed. The table, though empty of data, continues to exist until you use a DROP TABLE statement. TRUNCATE TABLE and DELETE without specifying any rows are functionally equivalent, but TRUNCATE TABLE is faster. DELETE removes rows one at a time and logs these transactions; TRUNCATE TABLE removes whole data pages and is not logged. Both DELETE and TRUNCATE TABLE reclaim the space occupied by the data and its associated indexes. The second FROM clause allows you to delete rows from one table based on data stored in other tables. You can define a trigger that takes a specified action when a DELETE statement is executed on a specified table. Permissions DELETE permission defaults to the table or view owner, who can transfer it to other users. See Also CREATE TRIGGER, DROP TABLE, TRUNCATE TABLE DISK INIT ──────────────────────────────────────────────────────────────────────────── Function Reserves and formats physical storage for database devices. Database devices contain databases and transaction logs. DISK INIT also provides a physical filename and a logical filename for the database file. (You need not use DISK INIT on the master database device; it is initialized when SQL Server is installed.) Syntax DISK INIT NAME = "logical_name" , PHYSNAME = "physical_name" , VDEVNO = virtual_device_number , SIZE = number_of_blocks [, VSTART = virtual_address , CNTRLTYPE = controller_number] Examples disk init name = "user_file", physname = "f:\sqlservr\data\userdisk.dat", vdevno = 2, size = 5120 Options NAME The logical name of the database device. It must correspond to the rules for identifiers and must be enclosed in single or double quotation marks. This name is used in the CREATE DATABASE and ALTER DATABASE statements. Although a name longer than 30 characters is allowed, the name will be truncated to 30 characters. PHYSNAME The full pathname of the database device. It must follow the rules for the operating system pathnames. It must be enclosed in single or double quotation marks. VDEVNO The virtual device number. It must be unique among database devices associated with SQL Server. The device number 0 is reserved for the master database device. Legal numbers are between 1 and 9. To determine the virtual device number, look at the device_number column of sysdevices with sp_helpdevice, and use the next integer. SIZE The size of the database device in 2K blocks. If you are planning to use the new database device for the creation of a new database, the minimum SIZE is the size of the model database, 1024 2K blocks (2 megabytes). If you are initializing a database device that will contain a transaction log, SIZE can be as small as 512 2K blocks (1 megabyte). VSTART The starting virtual address or the starting offset in 2K blocks. The value for VSTART should be 0 (the default). Reset VSTART only if instructed to do so. CNTRLTYPE The disk controller. Its default value is zero. Reset CNTRLTYPE only if instructed to do so. Comments Execute DISK INIT once for each new database device. Each time DISK INIT is executed, a row is added to master..sysdevices. It is important to back up the master database with the DUMP DATABASE or DUMP TRANSACTION statement after each use of DISK INIT. This makes recovery easier and safer in case master is damaged. (If you add a database device with DISK INIT and fail to back up master, you may be able to recover the changes by using DISK REINIT and then stopping and restarting SQL Server.) User databases are assigned to database devices with the optional "ON database_file" clause of the CREATE DATABASE or ALTER DATABASE statement. A new database device does not automatically become part of the pool of default database storage. Assign default status to a database device with the sp_diskdefault system procedure. To put a database's transaction log (that is, the system table syslogs) on a different database device than the one on which the rest of the database is stored, name at least two database devices when you create the database, and then execute the sp_logdevice system procedure. You can also alter the database to a second database device and then run sp_logdevice. For a report on all database devices and dump devices on your system, execute the sp_helpdevice system procedure. Remove a database device with the sp_dropdevice system procedure. You must first drop all existing databases on that database device. After dropping a database device, you can create a new one with the same logical name (using DISK INIT) as long as you give it a different physical name and virtual device number. If you wish to use the same physical name and virtual device number, you must restart SQL Server. Permissions DISK INIT permission defaults to the System Administrator and is not transferable. The System Administrator must be using the master database to use DISK INIT. See Also CREATE DATABASE, DISK REFIT, DISK REINIT, DUMP DATABASE, DUMP TRANSACTION, LOAD DATABASE, LOAD TRANSACTION, sp_diskdefault, sp_helpdevice, sp_logdevice DISK REFIT ──────────────────────────────────────────────────────────────────────────── Function Restores a damaged master database if databases have been created or database space allocated since the last database or transaction log dump. This statement is used after DISK REINIT. Syntax DISK REFIT Examples disk refit Comments DISK REFIT is used after DISK REINIT. It uses information in sysdevices to rebuild sysusages and sysdatabases. Stop and then restart SQL Server after using DISK REFIT. For more information, see the SQL Server System Administrator's Guide. Permissions DISK REFIT permission defaults to the System Administrator and is not transferable. The System Administrator must be in the master database to use DISK REFIT. See Also DISK INIT, DISK REINIT, sp_addumpdevice, sp_helpdevice DISK REINIT ──────────────────────────────────────────────────────────────────────────── Function Part of the procedure that restores a damaged master database if database devices have been added since the last database or transaction log dump. Syntax DISK REINIT NAME = "logical_name", PHYSNAME = "physical_name" , VDEVNO = virtual_device_number , SIZE = number_of_blocks [, VSTART = virtual_address , CNTRLTYPE = controller_number] Examples disk reinit name = "user_disk", physname = "f:\sqlservr\data\userdisk.dat", vdevno = 2, size = 5120 Options NAME The logical name of the database device. It must correspond to the rules for identifiers and must be enclosed in single or double quotation marks. This name is used in the CREATE DATABASE and ALTER DATABASE statements. Although a name longer than 30 characters is allowed, the name will be truncated to 30 characters. PHYSNAME The full pathname of the database device. It must follow the rules for the operating system pathnames. It must be enclosed in single or double quotation marks. VDEVNO The virtual device number. It must be unique among database devices used by SQL Server. The device number 0 is reserved for the master database device. Legal numbers are between 1 and 49. SIZE The size of the database device in 2K blocks. The minimum usable size is 1024 2K blocks (2 megabytes). VSTART The starting virtual address or the starting offset in 2K blocks. The value for VSTART should be 0 (the default). Reset it only if instructed to do so. CNTRLTYPE The disk controller. Its default value is 0. Reset it only if instructed to do so. Comments DISK REINIT insures that master..sysdevices is correct if a master database has been damaged and database devices have been added since the last dump of master. DISK REINIT is very similar to DISK INIT but does not initialize the disk space. See the SQL Server System Administrator's Guide for more complete information on using DISK REINIT. Permissions DISK REINIT permission defaults to the System Administrator and is not transferable. The System Administrator must be in the master database to use DISK REINIT. See Also DISK INIT, DISK REFIT, sp_addumpdevice, sp_helpdevice DROP DATABASE ──────────────────────────────────────────────────────────────────────────── Function Removes one or more databases from SQL Server. Syntax DROP DATABASE database_name [, database_name...] Examples A. drop database publishing B. drop database publishing, newpubs Examples A and B remove databases and their contents. Comments Removing a database deletes the database and all its objects, frees its storage allocation, and erases references to it in the master database. You cannot drop a database that is in use (open for reading or writing by any user). A damaged database cannot be removed with DROP DATABASE. Use the DBCC DBREPAIR statement: dbcc dbrepair (database_name, dropdb) Permissions DROP DATABASE permission defaults to the Database Owner and cannot be transferred. The master database must be current at the time. (A database is made current with the USE statement.) See Also ALTER DATABASE, CREATE DATABASE, DBCC, USE, sp_changedbowner, sp_helpdb, sp_renamedb, sp_spaceused DROP DEFAULT ──────────────────────────────────────────────────────────────────────────── Function Removes a user-specified default. Syntax DROP DEFAULT [owner.]default_name [, [owner.]default_name...] Examples drop default datedefault This example removes user-created default datedefault. Options default_name The name of an existing default. You can find out what defaults exist by executing the sp_help system procedure. Comments A default cannot be dropped if it is currently bound to a column or user datatype. Use the sp_unbindefault system procedure to unbind the default before you drop it. See Chapter 2, "System Procedures," for details. You can bind a new default to a column or user datatype without unbinding its current default. The new default will override the old one. When you drop a default for a NULL column, NULL will from then on be inserted in that position when rows are added but no value for that column is entered. When you drop a default for a NOT NULL column, you'll get an error message when rows are added but no value for that column is explicitly entered. Permissions DROP DEFAULT permission defaults to the owner of the default and is not transferable. See Also CREATE DEFAULT, CREATE RULE, DROP RULE, sp_bindefault, sp_help, sp_helptext, sp_rename, sp_unbindefault DROP INDEX ──────────────────────────────────────────────────────────────────────────── Function Removes an index from the database. Syntax DROP INDEX table_name.index_name [, table_name.index_name...] Examples drop index authors.au_id_ind This example removes the index au_id_ind in the authors table. Options table_name The table in which the indexed column is located. index_name The index to be dropped. In TRANSACT-SQL, index names need not be unique in a database, though they must be unique within a table. Comments Once the statement is executed, you regain all of the space occupied by the index. This space can be used for any database objects. You can't use DROP INDEX on any of the system tables that exist in the master database and in user databases. To get information about what indexes exist on a table or view, use the sp_helpindex system procedure with the table name as parameter. Permissions DROP INDEX permission defaults to the index owner and is not transferable. See Also CREATE INDEX, CREATE TABLE, sp_helpindex, sp_spaceused DROP PROCEDURE ──────────────────────────────────────────────────────────────────────────── Function Removes user-created stored procedures from the current database. Syntax DROP PROCedure procedure_name [procedure_name...] Examples use pubs <execute> drop procedure byroyalty This example removes the stored procedure byroyalty. Comments You can drop stored procedures only from the current database. The existence of a stored procedure is checked on each call to that procedure. A procedure group (more than one procedure with the same name but with different ;number suffixes) can be dropped with a single DROP PROCEDURE statement. For example, if the procedures used with the application orders were named orderproc;1, orderproc;2, and so on, the following statement would drop the entire group: drop proc orderproc Once procedures have been grouped, individual procedures within the group cannot be dropped. For example, the following statement is not allowed: drop procedure orderproc;2 The sp_helptext system procedure displays the procedure definition (which is stored in syscomments). See Chapter 2, "System Procedures," for details. Permissions DROP PROCEDURE permission defaults to the procedure owner and is not transferable. See Also CREATE PROCEDURE, EXECUTE, sp_depends, sp_helptext, sp_rename DROP RULE ──────────────────────────────────────────────────────────────────────────── Function Removes a user-specified rule. Syntax DROP RULE [owner.]rule_name [, [owner.]rule_name...] Examples drop rule pubid_rule This example removes the pubid_rule rule. Comments Use the sp_unbindrule system procedure to unbind the rule before dropping it. If it is still in effect, an error message will be displayed and the DROP RULE statement will be aborted. See Chapter 2, "System Procedures," for details. You can bind a new rule to a column or user datatype without unbinding its current rule. The new rule overrides the old one. After you drop a rule, new data entered into the columns that previously were governed by it goes in without these constraints. Already existing data is not affected in any way. Permissions DROP RULE permission defaults to the rule owner and is not transferable. See Also CREATE DEFAULT, CREATE RULE, DROP DEFAULT, sp_bindrule, sp_help, sp_helptext, sp_rename, sp_unbindrule DROP TABLE ──────────────────────────────────────────────────────────────────────────── Function Removes a table definition and all data, indexes, triggers, and permission specifications for a table from the database. Syntax DROP TABLE [[database.]owner.]table_name [, [[database.]owner.]table_name...] Examples drop table roysched This example removes the roysched table and its data and indexes. Comments When you drop a table, any rules or defaults on it lose their binding, and any triggers associated with it are automatically dropped. If you re-create a table, you must rebind the appropriate rules and defaults, and re-create any triggers. The system tables affected when a table is dropped are sysobjects, syscolumns, sysindexes, sysprotects, and syscomments. You can't use the DROP TABLE statement on any of the system tables, either in the master database or in a user database. You can drop a table in any database if you are the table owner. The following examples drop newtable in the otherdb database: drop table otherdb..newtable drop table otherdb.yourname.newtable If you delete all the rows in a table or give the TRUNCATE TABLE statement, the table still exists until you drop it. Permissions DROP TABLE permission defaults to the table owner and is not transferable. See Also ALTER TABLE, CREATE TABLE, DELETE, TRUNCATE TABLE, sp_depends, sp_help, sp_rename, sp_spaceused DROP TRIGGER ──────────────────────────────────────────────────────────────────────────── Function Removes a trigger. Syntax DROP TRIGGER [owner.]trigger_name [, [owner.]trigger_name...] Examples drop trigger trigger1 Permissions DROP TRIGGER permission defaults to the trigger owner and is not transferable. See Also CREATE TRIGGER, sp_depends, sp_help, sp_helptext, sp_rename DROP VIEW ──────────────────────────────────────────────────────────────────────────── Function Removes views from the database. Syntax DROP VIEW [owner.]view_name [, [owner.]view_name...] Examples drop view new_pricey This example removes the new_pricey view. Comments When you use DROP VIEW, the definition of the view and other information about it are deleted from the system tables sysobjects, syscolumns, syscomments, sysprocedures, and sysprotects. All permissions for the view are deleted, too. Existence of a view is checked each time the view is referenced by another view or by a stored procedure. Permissions DROP VIEW permission defaults to the view owner and is not transferable. See Also CREATE VIEW, Views, sp_depends, sp_help, sp_helptext, sp_rename DUMP DATABASE ──────────────────────────────────────────────────────────────────────────── Function Makes a backup copy of the database and the transaction log in a form that can be read in with LOAD DATABASE. Syntax DUMP DATABASE database_name TO dump_device Examples dump database pubs to diskdump1 This example dumps the pubs database to the diskdump1 diskette. Options database_name The name of the database from which you are copying data. TO dump_device The logical name of the dump device to which you are dumping the specified database. Logical dump device names are stored in master..sysdevices.name. A report on dump devices is displayed when you execute sp_helpdevice. (Logical dump device names are in the report column device_name.) You add dump devices with the sp_addumpdevice system procedure. See Chapter 2, "System Procedures," or the SQL Server System Administrator's Guide for details. Comments SQL Server database dump is a dynamic dump─it can take place while the database is active. It slows the system down minimally, so you may want to run it when the database is not being heavily updated. The dump device is on the machine on which SQL Server is running. When the DUMP DATABASE statement is executed for dumping to a diskette, the console program prompts the operator and waits for responses. (See Chapter 3, "Utility Programs," for more information on console.) If console is not running or if the operator does not answer its questions, the dump cannot proceed. The dump captures the state of the database as it was at the moment the statement was executed. Any changes made after the dump begins are not reflected in the dumped database. To recover a database, load the most recent database dump and load all the transaction log dumps made since that database dump. The database into which you are loading the data is replaced by this reconstruction. DUMP DATABASE backs up the database and the log but does not remove the inactive portion of the log. DUMP TRANSACTION backs up the log and removes the inactive portion of it. The DUMP DATABASE statement must be used to back up both the database and its transaction log for small databases. Databases smaller than about 4 megabytes can store their transaction logs on the same database device as the rest of the database. In this situation, DUMP DATABASE is used to create backups, and DUMP TRANSACTION WITH TRUNCATE_ONLY can be used to purge the log of completed transactions. Databases larger than 4 megabytes should store their transaction logs on a different database device, which improves performance and enables unrestricted use of both the DUMP DATABASE and DUMP TRANSACTION statements. You can run more than one dump (or load) at the same time, but only if the cntrltype values of the devices to which you are dumping or loading are different. For example, only one dump or load through the disk byte stream interface can be active at a time. For more information on SQL Server's backup and recovery facilities, see the SQL Server System Administrator's Guide. Permissions DUMP DATABASE permission defaults to the Database Owner, who can transfer it to other users. See Also DUMP TRANSACTION, LOAD DATABASE, LOAD TRANSACTION, sp_addumpdevice, sp_dropdevice, sp_helpdb, sp_helpdevice, sp_logdevice, sp_spaceused DUMP TRANSACTION ──────────────────────────────────────────────────────────────────────────── Function Removes the inactive part of the transaction log and makes a backup copy of it in a form that can be read in with LOAD TRANSACTION. The DUMP TRANSACTION and LOAD TRANSACTION statements should be used only when the transaction log is stored on a different database device than the rest of the database. Syntax DUMP TRANsaction database_name [TO dump_device] [WITH TRUNCATE_ONLY] [WITH NO_LOG] Examples dump transaction pubs to diskdump1 This example dumps the transaction log to the diskdump1 dump device. Options TO dump_device The logical name of the dump device to which you are dumping the specified transaction log. Logical dump device names are available in master..sysdevices.name. To add a dump device, use the sp_addumpdevice system procedure. The dump device name is optional if you use the WITH TRUNCATE_ONLY or WITH NO_LOG clause. WITH TRUNCATE_ONLY Removes the inactive part of the log without making a backup copy of it. This optional clause frees disk space used by the transaction log. WITH TRUNCATE_ONLY is used after you have backed up the entire database with DUMP DATABASE. The DUMP DATABASE statement backs up the log but does not remove the inactive portion of it. If you use WITH TRUNCATE_ONLY and do not have a backup created by DUMP DATABASE, the changes that had been recorded in the log are not recoverable. Since the WITH TRUNCATE_ONLY option does not perform a dump, you can use any dump device name. You cannot use both WITH TRUNCATE_ONLY and WITH NO_LOG in the same statement. WITH NO_LOG Is used only when you have run out of space in the database so that you cannot even run DUMP TRANSACTION WITH TRUNCATE_ONLY to retrieve some space from the log. Like WITH TRUNCATE_ONLY, WITH NO_LOG removes the inactive part of the log without making a backup copy of it. In addition, WITH NO_LOG saves space by not recording this procedure in the transaction log. After the transaction log has been dumped using WITH NO_LOG, the changes that had been recorded in the log are not recoverable. You should immediately run DUMP DATABASE and then enlarge the database with ALTER DATABASE. You cannot use both WITH TRUNCATE_ONLY and WITH NO_LOG in the same statement. Comments A transaction log dump can take place while the database is active. The DUMP TRANSACTION (and LOAD TRANSACTION) statement should be used only when the transaction log is stored on a different database file than the rest of the database. This is recommended for databases larger than about 4 megabytes. The transaction log is transferred to a separate database device with the sp_logdevice system procedure. With smaller databases, the transaction log is usually stored on the same database device as the rest of the database. In this situation, you use DUMP DATABASE to back up both the database and the transaction log, after which you can use DUMP TRANSACTION WITH TRUNCATE_ONLY to remove committed transactions from the log. DUMP TRANSACTION uses much less storage and takes much less time than DUMP DATABASE, which backs up both the transaction log and the rest of the database. When the DUMP TRANSACTION statement is executed for dumping to a diskette, the console program prompts the operator and waits for a response. If console is not running or if the operator does not answer its questions, the dump cannot proceed. Dumps of the transaction log are typically coordinated with DUMP DATABASE as part of an overall backup procedure. Transaction log dumps are made more frequently than database dumps. The dump captures the state of the transaction log as it was at the moment the statement was executed. Any changes made after the dump begins are not reflected in the dumped log. To recover a database, load the most recent database dump and load all the transaction log dumps (in order) made since the database dump. The database is restored to its state at the time of the last transaction log dump. ──────────────────────────────────────────────────────────────────────────── WARNING Do not attempt to modify the transaction log table syslogs with a DELETE, UPDATE, or INSERT statement. Doing so makes it impossible for SQL Server to recover correctly in case of a system failure. In addition, an attempt to delete all rows from syslogs causes SQL Server to get into an infinite loop that eventually fills up the entire database. ──────────────────────────────────────────────────────────────────────────── You can run more than one dump (or load) at the same time, but only if the cntrltype values of the devices to which you are dumping or loading are different. For example, only one dump or load through the disk byte stream interface can be active at a time. You cannot execute the DUMP TRANSACTION statement while the select into/bulkcopy or the trunc. log on chkpt. database options are enabled. Use DUMP DATABASE instead, or disable the option with sp_dboption. Permissions DUMP TRANSACTION permission defaults to the Database Owner, who can transfer it to other users. For more information on SQL Server's backup and recovery facilities, see the SQL Server System Administrator's Guide. See Also DUMP DATABASE, LOAD DATABASE, LOAD TRANSACTION, sp_addumpdevice, sp_dboption, sp_dropdevice, sp_helpdevice, sp_logdevice EXECUTE ──────────────────────────────────────────────────────────────────────────── Function Runs a system procedure or a user-defined stored procedure. This statement is only necessary if other statements are submitted in the same batch. Syntax [EXECute] [[database.]owner.]procedure_name[;number] [[@parameter_name =] value [,[@parameter_name =] value]...] [WITH RECOMPILE] Examples Supply parameters in either of two ways: by value or with "parameter_name = value". If you use the "parameter_name = value" form, you need not supply the parameters in the order defined in the CREATE PROCEDURE statement. A. execute showind titles B. exec showind @tabname = titles If this is the only statement in a batch or file, you can use the following form: C. showind titles The showind stored procedure with a parameter value titles has been executed. Options procedure_name The name of a procedure that has been defined with a CREATE PROCEDURE statement. You can execute a procedure that has been created in another database as long as you are its owner or have permission to execute it in that database. ;number An optional integer used to group procedures of the same name so that they can be dropped together with a single DROP PROCEDURE statement. Procedures used in the same application are often grouped this way. For example, the procedures used with the application orders might be named orderproc;1, orderproc;2, and so on. The statement DROP PROCEDURE orderproc would drop the entire group. Once procedures have been grouped, individual procedures within the group cannot be dropped. For example, the statement DROP PROCEDURE orderproc;2 is not allowed. parameter_name The name of a parameter to the procedure, as defined in the CREATE PROCEDURE statement. Parameter names must be preceded by the "at" symbol (value" form, parameter names and constants need not be supplied in the order defined in the CREATE PROCEDURE statement. value The value of the parameter or parameter to the procedure. If you do not use parameter names, parameter values must be supplied in the order defined in the CREATE PROCEDURE statement. If a default is defined in the CREATE PROCEDURE statement, a user can execute the procedure without giving a parameter. The default must be a constant. It can include the wildcard characters (%, _, [ ], and [^]) if the procedure uses the parameter name with the LIKE keyword. See example B in the "CREATE PROCEDURE" section. The default can be NULL. Usually, the procedure definition specifies what action should be taken if a parameter value is NULL. WITH RECOMPILE Forces compilation of a new plan only for this execution of the procedure. Use this option if the parameter you're supplying for this execution is atypical. The original plan is used on subsequent executions. Comments If the user does not supply expected parameters, the stored procedure error message lists the names of expected parameters in reverse order. You don't need to use the EXECUTE keyword if the statement is the first one in a batch. A batch is a segment of an input file terminated by the go command on a line by itself. Since stored procedures are compiled the first time they are executed, subsequent run time is much shorter than for the equivalent set of stand-alone statements. System procedures (for example, sp_help) do not work on temporary tables because a procedure cannot access a temporary object unless it created the object itself during the current session. See "System Procedures," later in this chapter, and Chapter 2, "System Procedures," for details. Permissions EXECUTE permission defaults to the owner of the stored procedure, who can transfer it to other users. See Also CREATE PROCEDURE, DROP PROCEDURE, Parameters, System Procedures, Variables (Local and Global), Wildcard Characters, sp_depends, sp_helptext Expressions ──────────────────────────────────────────────────────────────────────────── Function Used as variables and constants in many SQL statements. Syntax An expression returns values. Expressions have the following syntax: {constant | column_name | function | (subquery)} [{arithmetic_operator | bitwise_operator | string_operator} {constant | column_name | function | (subquery)}...] A boolean expression returns true or false. It has the following syntax: - expression comparison_operator [ANY | ALL] expression - expression [NOT] IN expression - [NOT] EXISTS expression - expression [NOT] BETWEEN expression AND expression - expression [NOT] LIKE expression - NOT expression LIKE expression - expression IS [NOT] NULL - NOT boolean_expression - boolean_expression {AND | OR} boolean_expression - [NOT] boolean_function Options function An aggregate function or a built-in function. See "Functions" for more information. subquery A nested SELECT statement with some restrictions. A subquery can be used in place of an expression if it returns a single value. See "Subqueries" for more information. arithmetic_operator One of the following: ╓┌───────────────────────┌───────────────────────────────────────────────────╖ Symbol Meaning Symbol Meaning ──────────────────────────────────────────────────────────────────────────── ++ Addition - Subtraction * Multiplication / Division % Modulo ──────────────────────────────────────────────────────────────────────────── Addition, subtraction, division, and multiplication can be used on int, smallint, tinyint, float, and money columns. The modulo operator cannot be used on money or float columns. Modulo finds the integer remainder after dividing two whole numbers. For example, 21 % 11 = 10, because 21 divided by 11 equals 1 with a remainder of 10. bitwise_operator One of the following: Symbol Meaning ──────────────────────────────────────────────────────────────────────────── & Bitwise AND (two operands) | Bitwise OR (two operands) Bitwise exclusive OR (two operands) ~ Bitwise NOT (one operand) ──────────────────────────────────────────────────────────────────────────── The bitwise operators that require two operands (&, |, and ^) can be used on columns with the datatypes shown below: ╓┌─────────────────┌─────────────────────────────────────────────────────────╖ Left Operand Right Operand ──────────────────────────────────────────────────────────────────────────── int int, smallint, tinyint, binary, or varbinary smallint int, smallint, tinyint, binary, or varbinary tinyint int, smallint, tinyint, binary, or varbinary varbinary int, smallint, or tinyint bit int, smallint, tinyint, or bit float none binary int, smallint, or tinyint ──────────────────────────────────────────────────────────────────────────── The ~ bitwise operator can be used on columns with a datatype of int, smallint, tinyint, or bit. All the bitwise operators translate the integer parameters into binary representations before evaluating them. In the following examples, A = 10101010 and B = 01001011: (A & B) 10101010 01001011 -------- 00001010 The result is 1 where both A and B are 1. (A | B) 10101010 01001011 -------- 11101011 The result is 1 where A, B, or A and B are 1. (A ^ B) 10101010 01001011 -------- 11100001 The result is 1 where A or B, but not both, are 1. (~A) 10101010 -------- 01010101 All 1s are changed to 0s and all 0s to 1s. string_operator The string operator + can be used to concatenate expressions. See "String Functions" for examples. comparison_operator One of the following: ╓┌────────────────────┌──────────────────────────────────────────────────────╖ Symbol Meaning ──────────────────────────────────────────────────────────────────────────── = Equal to > Greater than < Less than >= Greater than or equal to <= Less than or equal to != Not equal to !> Not greater than !< Not less than ──────────────────────────────────────────────────────────────────────────── In comparing char and varchar data, < means closer to the beginning of the alphabet and > means closer to the end of the alphabet. In the ASCII collating sequence, lowercase letters are greater than uppercase letters, and uppercase letters are greater than numbers. In comparing dates, < means earlier and > means later. Trailing blanks are ignored for comparison purposes. For example, "Dirk" is the same as "Dirk ". Put single or double quotation marks around all character and datetime data used with a comparison operator (= "Bennet", > "94609"). ANY Is used with <, >, or = and a subquery. It returns results when any value retrieved in the subquery matches the value in the WHERE or HAVING clause of the outer statement. (See "Subqueries" for more information.) ALL Is used with < or > and a subquery. It returns results when all values retrieved in the subquery match the value in the WHERE or HAVING clause of the outer statement. (See "Subqueries" for more information.) NOT Negates the meaning of a keyword, a boolean function, or a boolean expression. IN Returns results when any value returned by the second expression matches the value in the first expression. The second expression must be a subquery or a list of values enclosed in parentheses. IN is equivalent to = ANY. See "WHERE Clause" for details. EXISTS Is used to test for the existence of some result. The expression that follows must be a subquery. BETWEEN The range-start keyword. Use AND for the range-end keyword. A range of between x and y is inclusive. A range of > x < y is not inclusive. LIKE Indicates that the following expression is a matching pattern. With LIKE, the second expression must always be a character string. LIKE is only available for char, varchar, and datetime columns (but not to search for seconds or milliseconds). It can be used with these wildcards: Wildcard Meaning ──────────────────────────────────────────────────────────────────────────── % Any string of 0 or more characters _ (underscore) Any single character [ ] Any single character within the specified range ([a-f]) or set ([abcdef]) [^] Any single character not within the specified range ([^a-f]) or set ([^abcdef]) ──────────────────────────────────────────────────────────────────────────── The wildcard and the string must be enclosed in single or double quotation marks (LIKE "[dD]eFr_nce"). For complete information, see "Wildcard Characters." NULL Indicates the value is "unknown" to the system. Use IS NULL or IS NOT NULL in queries on columns defined as NULL. If you use IS NULL or IS NOT NULL to query columns defined as NOT NULL, SQL Server displays an error message. An expression with a bitwise or arithmetic operator evaluates to NULL if any of the operands is null. AND Connects two expressions and returns results when both are true. When more than one logical operator is used in a statement, AND is handled first. However, you can change the order of execution with parentheses. OR Connects two or more conditions and returns results when either of the conditions is true. When more than one logical operator is used in a statement, OR is evaluated after AND. However, you can change the order of execution with parentheses. boolean_function A function that returns true or false. Comments Parentheses can be used to group the elements in an expression or a boolean expression. When "expression" is given as a variable in a syntax statement, a simple expression is assumed. "Boolean expression" is specified when only a boolean expression is acceptable. Operators have the following precedence levels, where 1 is the highest level and 6 is the lowest level: 1. (unary) ~ - + 2. * / % 3. + - & | ^ 4. NOT 5. AND 6. OR When all operators in an expression are of the same level, the order of execution is left to right. You can change the order of execution with parentheses─the most deeply nested expression is evaluated first. There are two ways to specify literal quotation marks within a char or varchar entry. The first method is to use two quotation marks. For example, if you have begun a character entry with a single quotation mark and wish to include a single quotation mark as part of the entry, use two single quotation marks: 'I don"t understand.' With double quotation marks: "He said, ""It's not really confusing.""" The second method is to enclose a quote in the opposite kind of quotation mark. In other words, surround an entry containing a double quotation mark with single quotation marks (or vice versa). Here are some examples: 'George said, "There must be a better way."' "Isn't there a better way?" 'George asked, "Isn"t there a better way?"' To continue a character string to the next line on your screen, enter a backslash (\) before going to the next line. See Also Functions, Joins, Search Conditions, SELECT, Subqueries, WHERE Clause, Wildcard Characters, sp_helpjoins Functions ──────────────────────────────────────────────────────────────────────────── Function Return special information from the database. Functions are used in SELECT statements and are divided into aggregate functions (SUM, AVG, COUNT, COUNT(*), MAX, and MIN), date functions, mathematical functions, string functions, system functions, text/image functions, and type conversion functions. Each of these groups is discussed in its own section of this manual. Syntax Aggregate functions return summary values. See "Aggregate Functions" for syntax and other information. Date functions compute datetime values and their components, dateparts. See "Date Functions" for syntax and other information. Mathematical functions perform operations on numeric data. See "Mathematical Functions" for syntax and other information. String functions perform operations on character and binary data. See "String Functions" for syntax and other information. System functions return special information from the database. See "System Functions" for syntax and other information. Text/image functions perform operations on text and image data. See "Text/Image Functions" for syntax and other information. Conversion functions convert expressions of one datatype to another datatype. See "Conversion Function" for syntax and other information. Comments Unlike the aggregate functions, the other built-in functions are often used on their own or as part of a stored procedure or program. When they are used in SQL statements, they are allowed in the select list, in the WHERE clause, and anywhere an expression is allowed. Many of the miscellaneous built-in functions are SQL Server extensions to SQL. See Also Aggregate Functions, COMPUTE Clause, Conversion Function, Date Functions, Mathematical Functions, Search Conditions, SELECT, String Functions, System Functions, Text/Image Functions GOTO ──────────────────────────────────────────────────────────────────────────── Function Causes an unconditional branching to a user-defined label. Syntax label: GOTO label Examples declare @count smallint select @count = 1 restart: print "yes" select @count = @count + 1 while @count <= 4 goto restart Comments The label name must follow the rules for identifiers and must be followed by a colon (:) when it is declared. It is not followed by a colon when it is used with GOTO. GOTO is usually made dependent on an IF or WHILE test, or some other condition to avoid an endless loop between GOTO and the label. See Also Control-of-Flow Language, IF...ELSE, WHILE GRANT ──────────────────────────────────────────────────────────────────────────── Function Assigns permissions to users. Syntax Object permissions: GRANT {ALL | permission_list} ON {table_name [(column_list)] | view_name [(column_list)] | stored_procedure_name} TO {PUBLIC | name_list} Statement permissions: GRANT {ALL | statement_list} TO {PUBLIC | name_list} Examples A. use pubs /* Permissions can be granted on objects in the ** current database only. */ grant insert, delete on titles to mary, sales B. grant update on titles (royalty, advance) to public C. grant create database, create table to mary, john D. grant all on titles to public E. grant all to public Options ALL When used to assign object permissions (first syntax format), this keyword specifies that all permissions applicable to the specified object are granted or revoked. Only the System Administrator can use ALL to assign statement permissions (second syntax format), since only the System Administrator can grant or revoke CREATE DATABASE permission. permission_list A list of permissions granted. When permissions are granted on a table or a view, the permission list can include one or more of the following items: SELECT, INSERT, DELETE, and UPDATE. When permissions are granted on columns, the permission list can include one or both of the following items: SELECT and UPDATE. When permissions are granted on stored procedures, the permission list can include EXECUTE only. If more than one permission is listed, separate them with commas. statement_list A list of statements granted. The statement list can include CREATE DATABASE (if the user executing the statement is the System Administrator), CREATE DEFAULT, CREATE PROCEDURE, CREATE RULE, CREATE TABLE, CREATE VIEW, DUMP DATABASE, and DUMP TRANSACTION. If more than one statement is listed, separate them with commas. table_name The name of a table in the current database. Only one table can be listed for each GRANT statement. column_list A list of columns, separated by commas, to which the permissions apply. If columns are specified, only SELECT and UPDATE permissions can be granted. view_name The name of a view in the current database. Only one view can be listed for each GRANT statement. stored_procedure_name The name of a stored procedure in the current database. Only one stored procedure can be listed for each GRANT statement. PUBLIC All users in the public group, which includes all users not explicitly in another group. name_list A list of users' database names and/or groupnames, separated by commas. These names are stored in the database's sysusers table. Users are added with sp_adduser; groups are added with sp_addgroup. New users are made members of groups with sp_adduser; group membership is modified with sp_changegroup. Comments You can substitute the word "FROM" for "TO" in the GRANT syntax. Table summarizes the SQL Server protection system. Permission to use each statement defaults to the System Administrator, the Database Owner, the table owner, or the default group public. This user can grant or revoke the permission if it is transferable. Users at higher levels are either automatically assigned the permission or (in the case of Database Owners) can get it via the SETUSER statement. For example, the owner of a database does not automatically receive permissions on objects owned by other users. A Database Owner can always take any permission by assuming the identity of the object owner with the SETUSER statement and then executing the appropriate GRANT or REVOKE statements. The System Administrator has permission to do anything at any time. For permissions that default to PUBLIC, no permission is required─that is, no GRANT or REVOKE statements must be written. Table Summary of Permissions System ╓┌────────────┌──────────┌──────────┌──────────┌───────┌──────────┌───┌──────╖ Can Be Defaults Granted / to Revoked System Db Owner Table Public Yes No N/A Admin Owner Can Be Defaults Granted / to Revoked System Db Owner Table Public Yes No N/A Admin Owner Statement: ──────────────────────────────────────────────────────────────────────────── ALTER X ─ ─ ─ (1) ─ ─ DATABASE ALTER TABLE ─ ─ X ─ ─ X ─ BEGIN ─ ─ ─ X ─ ─ X TRANSACTION CHECKPOINT ─ X ─ ─ ─ X ─ COMMIT ─ ─ ─ X ─ ─ X TRANSACTION CREATE X ─ ─ ─ X ─ ─ Can Be Defaults Granted / to Revoked System Db Owner Table Public Yes No N/A Admin Owner Statement: CREATE X ─ ─ ─ X ─ ─ DATABASE CREATE ─ X ─ ─ X ─ ─ DEFAULT CREATE ─ ─ X ─ ─ X ─ INDEX CREATE ─ X ─ ─ X ─ ─ PROCEDURE CREATE RULE ─ X ─ ─ X ─ ─ CREATE ─ X ─ (2) X(2) ─ ─ Can Be Defaults Granted / to Revoked System Db Owner Table Public Yes No N/A Admin Owner Statement: CREATE ─ X ─ (2) X(2) ─ ─ TABLE CREATE TRIG ─ ─ X ─ ─ X ─ GER CREATE VIEW ─ X ─ ─ X ─ ─ DBCC ─ X ─ ─ ─ ─ X DELETE ─ ─ X(3) ─ X ─ ─ DISK INIT X ─ ─ ─ ─ X ─ DISK REFIT X ─ ─ ─ ─ X ─ Can Be Defaults Granted / to Revoked System Db Owner Table Public Yes No N/A Admin Owner Statement: DISK REFIT X ─ ─ ─ ─ X ─ DISK REINIT X ─ ─ ─ ─ X ─ DROP (any ─ ─ ─ ─ ─ X ─ object)(4) DUMP ─ X ─ ─ X ─ ─ DATABASE DUMP ─ X ─ ─ X ─ ─ TRANSACTION EXECUTE (5) ─ ─ ─ ─ X ─ ─ Can Be Defaults Granted / to Revoked System Db Owner Table Public Yes No N/A Admin Owner Statement: GRANT ─ X ─ ─ ─ X ─ GRANT on ─ ─ ─ ─ ─ X ─ object (4) INSERT ─ ─ X(3) ─ X ─ ─ KILL X ─ ─ ─ ─ X ─ LOAD ─ X ─ ─ ─ X ─ DATABASE LOAD ─ X ─ ─ ─ X ─ TRANSACTION Can Be Defaults Granted / to Revoked System Db Owner Table Public Yes No N/A Admin Owner Statement: TRANSACTION PRINT ─ ─ ─ X ─ ─ X RAISERROR ─ ─ ─ X ─ ─ X READTEXT ─ ─ X(3) ─ X ─ ─ RECONFIGURE X ─ ─ ─ ─ X ─ REVOKE ─ X ─ ─ ─ X ─ REVOKE on ─ ─ ─ ─ ─ X ─ object (4) Can Be Defaults Granted / to Revoked System Db Owner Table Public Yes No N/A Admin Owner Statement: ROLLBACK ─ ─ ─ X ─ ─ X TRANSACTION SAVE ─ ─ ─ X ─ ─ X TRANSACTION SELECT ─ ─ X(3) ─ X ─ ─ SET ─ ─ ─ X ─ ─ X SETUSER ─ X ─ ─ ─ X ─ TRUNCATE ─ ─ X ─ ─ X ─ TABLE Can Be Defaults Granted / to Revoked System Db Owner Table Public Yes No N/A Admin Owner Statement: TABLE UPDATE ─ ─ X(3) ─ X ─ ─ UPDATE ─ ─ X ─ ─ X ─ STATISTICS WRITETEXT ─ ─ X(3) ─ X ─ ─ ──────────────────────────────────────────────────────────────────────────── (1) Transferred with CREATE DATABASE permission (2) Public can create temporary tables, no permission required (3) If a view, permission defaults to view owner (4) Defaults to object owner (5) Defaults to stored procedure owner You can grant or revoke permissions on objects in the current database only. The statement permissions include CREATE DATABASE (which can be granted only by the System Administrator), CREATE DEFAULT, CREATE PROCEDURE, CREATE RULE, CREATE TABLE, CREATE VIEW, DUMP DATABASE, and DUMP TRANSACTION. User groups allow you to grant or revoke permissions to more than one user with a single statement. Each user can be a member of at most one group. The Database Owner or System Administrator can add new users with sp_adduser and create groups with sp_addgroup. A guest user can be added with sp_adduser to allow users with login IDs on SQL Server to use the database with limited permissions. By default, all users are a member of the group public if they have not explicitly been added to another group. To add a new user to a group other than public, use the sp_adduser system procedure. To change an established user's group, use the sp_changegroup system procedure. To remove a user, use the sp_dropuser system procedure. To remove a group, use sp_dropgroup. To display the members of a group, use sp_helpgroup. Its optional parameter is the groupname. If you do not give a groupname, sp_helpgroup displays all the groups that exist and lists the members of each. The sp_helprotect system procedure reports permissions on a database object or on a user. See Chapter 2, "System Procedures," for details on the procedures for adding and dropping groups and users. The SQL Server setup program assigns a set of permissions to the default group, public. GRANT and REVOKE statements are order sensitive. The statement that takes effect when there is a conflict is the most recently executed one. Permission to grant and revoke defaults to particular users and cannot be transferred. See Also REVOKE, SETUSER, sp_addgroup, sp_adduser, sp_changedbowner, sp_changegroup, sp_dropgroup, sp_dropuser, sp_helpgroup, sp_helprotect, sp_helpuser GROUP BY and HAVING Clauses ──────────────────────────────────────────────────────────────────────────── Function Divide a table into groups. These clauses are used in SELECT statements. Syntax GROUP BY [ALL] aggregate_free_expression [, aggregate_free_expression...] [HAVING search_conditions] Examples A. select type, avg(advance), sum(ytd_sales) from titles group by type Example A calculates the average advance and the sum of the sales for each type of book. B. select type, pub_id, avg(advance), sum(ytd_sales) from titles group by type, pub_id Example B groups the results by type and then by pub_id within each type. C. select type, avg(price) from titles group by type having type like 'p%' Example C displays results for groups having a type beginning with "p" only. D. select pub_id, sum(advance), avg(price) from titles group by pub_id having sum(advance) > $15000 and avg(price) < $10 and pub_id > "0800" Example D displays results for groups matching the conditions in the HAVING clause. Options GROUP BY Specifies the groups into which the table will be divided and, if aggregate functions are included in the select list, finds a summary value for each group. These summary values appear as new columns in the results, rather than as new rows. When GROUP BY is used with standard SQL, each item in the select list must either have a fixed value in every row in the group or be used with aggregate functions that produce a single value for each group. TRANSACT-SQL has no such restrictions on the items in the select list. Also, TRANSACT-SQL allows you to group by a column name or any expression (except a column heading or alias), while with standard SQL you can only group by a column name. You can use these aggregate functions with GROUP BY (the expression in the syntax is usually a column name): ╓┌─────────────────────────────────┌─────────────────────────────────────────╖ Aggregate Function Result ──────────────────────────────────────────────────────────────────────────── SUM([DISTINCT] expression) Returns the total of the [distinct] values in the numeric column Aggregate Function Result ──────────────────────────────────────────────────────────────────────────── AVG([DISTINCT] expression) Returns the average of the [distinct] values in the numeric column COUNT([DISTINCT] expression) Returns the number of [distinct] non-null values in the expression COUNT(*) Returns the number of selected rows MAX(expression) Returns the highest value in the expression MIN(expression) Returns the lowest value in the expression ──────────────────────────────────────────────────────────────────────────── A table can be grouped by any combination of columns─that is, groups can be nested within each other. ALL Includes all groups in the results, even those that don't have any rows that meet the search conditions. For example: select type, avg(price) from titles where royalty = 10 group by all type ╓┌─┌──────────────────┌──────────────────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type ------------ ---------- UNDECIDED NULL business 17.31 mod_cook NULL popular_comp 20.00 psychology 14.14 ──────────────────────────────────────────────────────────────────────────── psychology 14.14 trad_cook 17.97 (6 rows affected) ALL is meaningful only if the SELECT statement includes a WHERE clause. aggregate_free_expression An expression that includes no aggregate functions. TRANSACT-SQL allows you to group by such an expression as well as by a column name. Note that you cannot group by a column heading or alias. This example is correct: select Price=avg(price), Pay=avg(advance), Total=price * $1.15 from titles group by price * $1.15 HAVING Sets conditions for the GROUP BY clause, similar to the way that WHERE sets conditions for the SELECT clause. HAVING search conditions are identical to WHERE search conditions with one exception: WHERE search conditions cannot include aggregate functions and HAVING search conditions can. Here's an example of a HAVING clause with aggregate functions: select pub_id, total = sum(ytd_sales) from titles where ytd_sales is not null group by pub_id having count(*)>5 There is no limit on the number of conditions that can be included in a HAVING clause. HAVING can be used without GROUP BY. If there are columns in the select list that neither have aggregate functions applied to them nor are included in the query's GROUP BY clause (illegal in standard SQL), the meanings of HAVING and WHERE are somewhat different. For examples, see the next section, "Comments." Comments You can use a column name or any expression (except a column heading) after GROUP BY. Null values in the GROUP BY column are put into a single group. Columns of type text or image are invalid in GROUP BY and HAVING clauses. Aggregate functions can be used in the select list or in the HAVING clause of a SELECT statement. They cannot be used in a WHERE clause. The aggregate functions, which calculate summary values from the non-null values in a column, can be divided into two groups. If they are applied to all the rows in a table (producing a single value per function), they are called scalar aggregate functions. If they are applied to all the rows that have the same value in a specified column or expression with the GROUP BY and, optionally, the HAVING clause (producing a value for each group per function), they are called vector aggregate functions. In either case, the results of the aggregate functions are shown as new columns. You can nest a vector aggregate function inside a scalar aggregate function. (See "Aggregate Functions" for more information and an example.) In standard SQL, columns in a select list that include aggregate functions must either have aggregate functions applied to them or be in the GROUP BY list. TRANSACT-SQL has no such restrictions. The first example shows a SELECT statement with the standard restrictions; the second one shows the same statement with another item (title_id) added to the select list to illustrate the difference in displays. The "extra" columns can also be referenced in the HAVING clause. For example: A. select type, avg(price) from titles group by type ╓┌─┌──────────────────┌──────────────────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type --------------- ----------- UNDECIDED NULL ──────────────────────────────────────────────────────────────────────────── UNDECIDED NULL business 13.73 mod_cook 11.49 popular_comp 21.48 psychology 13.50 trad_cook 15.96 (6 rows affected) B. select type, title_id, avg(price) from titles group by type ╓┌─┌───────────────────┌─────────┌───────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type title_id --------------- -------- --------- business BU1032 13.73 business BU1111 13.73 business BU2075 13.73 ──────────────────────────────────────────────────────────────────────────── business BU2075 13.73 business BU7832 13.73 mod_cook MC2222 11.49 mod_cook MC3021 11.49 UNDECIDED MC3026 NULL popular_comp PC1035 21.48 popular_comp PC8888 21.48 popular_comp PC9999 21.48 psychology PS1372 13.50 psychology PS2091 13.50 psychology PS2106 13.50 psychology PS3333 13.50 psychology PS7777 13.50 trad_cook TC3218 15.96 trad_cook TC4203 15.96 trad_cook TC7777 15.96 (18 rows affected) If there are columns in the select list that neither have aggregate functions applied to them nor are included in the query's GROUP BY clause (illegal in standard SQL), the meanings of HAVING and WHERE are somewhat different. In this situation, a WHERE clause restricts the rows that are included in the calculation of the aggregate function but does not restrict the rows returned by the query. Conversely, a HAVING clause restricts the rows returned by the query but does not affect the calculation of the aggregate function. The following examples illustrate the situation: A. select pub_id, count(pub_id) from publishers ╓┌─┌──────────────────┌──────────────────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── pub_id ----------- --------- 0736 3 0877 3 ──────────────────────────────────────────────────────────────────────────── 0877 3 1389 3 (3 rows affected) B. select pub_id, count(pub_id) from publishers where pub_id < "1000" ╓┌─┌──────────────────┌──────────────────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── pub_id ------------ ----------- 0736 2 0877 2 1389 2 (3 rows affected) Since the select list includes the pub_id column, which has no aggregate function applied and is not grouped, the WHERE clause returns all the values for pub_id, although it counts only rows where pub_id is less than 1000. C. select pub_id, count(pub_id) from publishers having pub_id < "1000" ╓┌─┌──────────────────┌──────────────────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── pub_id ------------ --------- 0736 3 0877 3 (2 rows affected) In example C, the HAVING clause means that only rows where pub_id is less than 1000 are returned. All the rows in the table are counted for the purposes of calculating the aggregate function. D. select pub_id, count(pub_id) from publishers where pub_id < "1000" having pub_id < "1000" ──────────────────────────────────────────────────────────────────────────── pub_id ------------ -------- 0736 2 0877 2 (2 rows affected) Since both HAVING and WHERE clauses are included (example D), only rows where pub_id is less than 1000 are returned, and only these rows are included in the calculation of the COUNT aggregate function. See Also Aggregate Functions, COMPUTE Clause, Functions, Row Aggregate Functions, Search Conditions, SELECT, WHERE Clause Identifiers ──────────────────────────────────────────────────────────────────────────── Function Name database objects─databases, tables, views, columns, indexes, triggers, procedures, defaults, rules, and so on. Syntax Identifiers are 1 to 30 characters long. The first character must be a letter or the symbols #, or _. (A table name beginning with # denotes a temporary table.) Characters following the first character can include letters, digits, or the symbols #, $, or _. No embedded spaces are allowed in identifiers. ──────────────────────────────────────────────────────────────────────────── NOTE When an identifier is preceded by the "at" symbol (@), the identifier is from 1 to 29 characters long. (For example, variable names and stored procedure parameter names are preceded by "@".) ──────────────────────────────────────────────────────────────────────────── By default, SQL Server is not sensitive to the case (upper or lower) of identifiers and data. If required, a System Administrator can build a case-sensitive SQL Server with the bldmastr program. The case of switches for utility programs is significant. Object names need not be unique in a database. However, column names and index names must be unique within a table, and other object names must be unique for each owner within a database. Database names must be unique on SQL Server. You can uniquely identify a table or column by adding other names that qualify it─the database name, owner's name, and (for a column) the table or view name. Each of these qualifiers is separated from the next by a period: database.owner.table_name.column_name database.owner.view_name.column_name The naming conventions are indicated in the syntax as follows: [[database.]owner.]table_name [[database.]owner.]view_name The default value for owner is the current user, and the default value for database is the current database. If desired, intermediate elements in a name can be omitted and their positions indicated by periods, as long as the system is given enough information to identify the object: database..table_name database..view_name Comments When qualifying a column name and a table name in the same statement, be sure to use the same qualifying expressions for each; they are evaluated as strings and must match or an error is returned. Two examples are shown, with different entries for the column name. Example B is not correct because the syntax for the column name does not match the syntax for the table name. A. select demo.mary.publishers.city from demo.mary.publishers city ----------------------- Boston Washington Berkeley B. select demo.mary.publishers.city from demo..publishers The column prefix "demo.mary.publishers" does not match a table name or alias name used in the query. The names of system datatypes and built-in functions are not reserved words and may be used to name database objects. When you reference an object without qualifying it with the database name and owner name, SQL Server tries to find the object in the current database among the objects you own. You need not qualify references to your objects in the current database, nor do you need to qualify objects owned by the Database Owner. However, you must qualify objects owned by other users by using the user's name. You can rename user objects (including user-defined datatypes) with sp_rename. See Also CREATE DATABASE, CREATE DEFAULT, CREATE PROCEDURE, CREATE RULE, CREATE TABLE, CREATE TRIGGER, CREATE VIEW, Datatypes, SELECT, sp_rename IF...ELSE ──────────────────────────────────────────────────────────────────────────── Function Impose conditions on the execution of an SQL statement. The SQL statement following an IF keyword and its condition executes if the condition is satisfied (when the boolean expression returns "true"). The optional ELSE keyword introduces an alternate SQL statement that executes when the IF condition is not satisfied (when the boolean expression returns "false"). Syntax IF boolean_expression statement [ELSE [boolean_expression] statement] Examples A. if 3 > 2 print "yes" B. if exists (select zip from authors where zip = "94705") print "Berkeley author" C. if (select max(id) from sysobjects) < 50 print "No user-created objects in this database" else begin print "These are the user-created objects" select name, type, id from sysobjects where id > 50 end The IF...ELSE condition tests for the presence of user-created objects (all of which have ID numbers that are larger than 50) in a database. Where user tables exist, the ELSE clause prints a message and selects their names, types, and ID numbers. Options boolean_expression An expression (a column name, a constant, any combination of column names and constants connected by arithmetic or bitwise operators, or a subquery) that returns "true" or "false." If the boolean expression contains a SELECT statement, the SELECT statement must be enclosed in parentheses. Comments The IF or ELSE condition can affect the performance of only a single SQL statement unless statements are grouped into a block between the keywords BEGIN and END, as in the last example. If a SELECT statement is used as part of the boolean expression, it must return a single value. The statement clause could be an EXECUTE statement or any other legal SQL statement or statement block. IF...ELSE constructs can be used either in a stored procedure (where they are often used to test for the existence of some parameter) or in ad hoc queries as in the earlier examples. IF tests can be nested, either within another IF or following an ELSE. There is no limit on the number of levels of nesting. See Also BEGIN...END, Control-of-Flow Language, CREATE PROCEDURE, Expressions, Subqueries INSERT ──────────────────────────────────────────────────────────────────────────── Function Adds new rows to a table or view. Syntax INSERT [INTO] [[database.]owner.]{table_name | view_name} [(column_list)] {VALUES (constant_expression [, constant_expression]...) | select_statement} Examples A. insert titles values("BU2222", "Faster!", "business", "1389", null, null, null, null, "ok", "06/17/87") B. insert titles(title_id, title, type, pub_id, notes, pubdate) values ('BU1237', 'Get Going!', 'business', '1389', 'great', '06/18/86') C. insert authors select * from newauthors where city = "San Francisco" D. insert test select * from test where city = "San Francisco" Options INTO An optional keyword. column_list A list of one or more columns to which data is to be added. The columns can be in any order, but the incoming data (whether in a VALUES clause or a SELECT clause) must be in the same order. The column_list is only necessary when some, but not all, of the columns in the table are to receive data. Enclose the items in column_list in parentheses. If no column_list is given, all of the columns in the receiving table (in CREATE TABLE order) is assumed. The column_list determines the order in which values are entered. In the first example, the columns in the column_list of the receiving table match the columns in the select list of the contributing table. Given a name like Euphemia Briggs, Euphemia (the first name) will be stored in the au_fname column and Briggs (the last name) will be stored in the au_lname column. In the second example, the order of the columns in the select list is not the same as the order of the columns in column_list. Euphemia (the first name) is stored in the au_lname column and Briggs (the last name) is stored in the au_fname column. A. insert authors (au_lname, au_fname) select au_lname, au_fname from newauthors B. insert authors (au_lname, au_fname) select au_fname, au_lname from newauthors VALUES Introduces a list of constant expressions. constant_expression Constant or null values for the indicated columns. The values list must be enclosed in parentheses and must match the explicit or implicit columns list. Enclose char, varchar, and datetime constants in single or double quotation marks. select_statement A standard SELECT statement used to retrieve the values to be inserted. Comments INSERT adds new rows only. Use UPDATE to modify any column values in a row you've already inserted. You can leave out items in the column_list and VALUES list as long as the omitted columns are defined to allow null values. See example B. You can select rows from a table and insert them into the same table in a single statement. See example D. If you have two tables identical in structure except that one has NULL fields and some NULL data and the other has NOT NULL fields, this difference makes it impossible to insert the data from the NULL table into the NOT NULL table with a SELECT statement. INSERT interacts with the IGNORE_DUP_KEY, IGNORE_DUP_ROW, and ALLOW_DUP_ROW options set with the CREATE INDEX statement. (See "CREATE INDEX" for more information.) A rule can be created and bound to a column to restrict the domain of legal values that can be entered into it. Rules are created with CREATE RULE and bound with the sp_bindrule system procedure. A default can be created and bound to a column in order to supply a value if a user does not explicitly enter one. Defaults are created with CREATE DEFAULT and bound with the sp_bindefault system procedure. You can define a trigger that takes a specified action when an INSERT statement is executed on a specified table. If an INSERT statement violates domain or integrity rules (see "CREATE RULE" and "CREATE TRIGGER") or if it is the wrong datatype (see "CREATE TABLE" and "Datatypes"), the statement fails, and SQL Server displays an error message. Permissions INSERT permission defaults to the table or view owner, who can transfer it to other users. See Also CREATE DEFAULT, CREATE INDEX, CREATE RULE, CREATE TABLE, CREATE TRIGGER, Datatypes, DELETE, SELECT, UPDATE, sp_bindefault, sp_bindrule, sp_unbindefault, sp_unbindrule Joins ──────────────────────────────────────────────────────────────────────────── Function Compare two or more tables (or views) by specifying a column from each, comparing the values in those columns row by row, and concatenating rows that have matching values. Syntax FROM {table_list | view_list} WHERE [NOT] [table_name. | view_name.] column_name join_operator [table_name. | view_name.] column_name [{AND | OR} [NOT] [table_name. | view_name.] column_name join_operator [table_name. | view_name.] column_name...] Examples A. select au_fname, au_lname, pub_name from authors, publishers where authors.city = publishers.city B. select au_fname, au_lname, pub_name from authors, publishers where authors.city *= publishers.city C. update titles set price = price * 2 from titles, publishers where titles.pub_id = publishers.pub_id and publishers.state = "CA" D. select type, au_lname, au_fname, title from authors, titles, titleauthor where authors.au_id = titleauthor.au_id and titles.title_id = titleauthor.title_id order by type E. select au_fname, au_lname, pub_name from authors, publishers where authors.city = publishers.city and authors.state = publishers.state F. select au1.au_fname, au1.au_lname, au2.au_fname, au2.au_lname from authors au1, authors au2 where au1.zip = au2.zip and (au1.au_lname + au1.au_fname) != (au2.au_lname + au2.au_fname) Example F finds the authors that live in the same zip code area as other authors, eliminating the rows where the author's name matches itself. G. select titles.type, t1.title_id, t1.price, t2.title_id, t2.price from titles t1, titles t2 where t1.type = t2.type and t1.price != t2.price Options FROM table_list, view_list The table(s) and view(s) that are being joined. All tables and/or views referenced elsewhere in the statement must be included in the FROM clause. WHERE clause The connection between the table(s) and view(s) named in the FROM clause, restricting the rows to be included in the results. Qualify column names if there is ambiguity about the table or view to which they belong. Columns being joined must be join compatible─they must both be numeric or both be character columns (or SQL Server must be instructed to treat them that way with the CONVERT function). join_operator One of the following: ╓┌─────────────────────────────────┌─────────────────────────────────────────╖ Symbol Meaning ──────────────────────────────────────────────────────────────────────────── = Equal to. > Greater than. < Less than. >= Greater than or equal to. <= Less than or equal to. != Not equal to. !> Not greater than. Symbol Meaning ──────────────────────────────────────────────────────────────────────────── !> Not greater than. !< Not less than. *= Includes all rows from the first table that meet the statement's restrictions. The second table generates values if there is a match on the join condition. Otherwise, the second table generates null values. =* Includes in the results all rows from the second table that meet the statement's restrictions. The first table generates values if there is a match on the join condition. Otherwise, the first table generates null values. ──────────────────────────────────────────────────────────────────────────── Symbol Meaning ──────────────────────────────────────────────────────────────────────────── ──────────────────────────────────────────────────────────────────────────── Comments If a join is to have meaningful results, the columns being joined must have similar values─values drawn from the same domain. You can use more than one join operator either to join more than two tables or to join more than two pairs of columns. (See examples D and E.) These "join expressions" are almost always connected with AND, although OR is also legal. The expression NOT column_name = column_name is equivalent to column_name != column_name. Joins based on equality (=) are called equijoins. A natural join is an equijoin with only one of the matching columns displayed in the results. Joins that compare values within the same column of one table are called self-joins. To distinguish the two roles in which the table appears, use aliases. See example F. Joins based on a comparison of scalar values (=, > , >= , < , <= , !=, !<, !>) are called theta joins. The "not equal join" is meaningful only with a self-join. See example G. Otherwise, it is rarely used, except by mistake. For example, it would be easy to think you could find the authors that do not live in a city in which a publisher is located with a not equal join: /* INCORRECT STATEMENT */ select distinct au_lname, authors.city from publishers, authors where publishers.city != authors.city However, this query actually finds the authors who live in a state where some publisher is not located, which is all of them. The correct SQL statement is as follows: select distinct au_lname, authors.city from publishers, authors where authors.city not in (select city from publishers where authors.city = publishers.city) Joins that include all rows regardless of whether there is a matching row (*= and =*) are called outer joins. See example B. If a table is an inner member of an outer join (the second table is the inner table if the join operator is *= ; the first table is the inner table if the join operator is =*), you can't place any restrictions on it except outer join restrictions. Null values in tables or views being joined will never match each other. Joins can also be stated as subqueries. (See the section on "Subqueries" for more information.) Joins may not be used for columns containing text or image values. See Also DELETE, INSERT, SELECT, Subqueries, WHERE Clause, UPDATE, sp_helpjoins KILL ──────────────────────────────────────────────────────────────────────────── Function Kills a process. Syntax KILL spid Examples kill 1378 Options spid The identification number of the process you want to kill, stored in sysprocesses.spid. The sysprocesses table also stores other process information, like status (runnable, waiting for a lock, infected, and so on), the uid (the ID of the user who executed the statement), program_name (the name of the application program), and so on. Comments You can get a report on the current processes with the sp_who system procedure. Here's a typical report: ╓┌───────┌───────────┌───────────┌───────────┌───────────┌───────┌───────┌─── ───────────────────────────────────────────────────────────────────────────── spid status login_id hostname blk dbname cmd ---- --------- -------- --------- --- ----- ----- 1 sleeping sa PC Client 0 pubs AWAIT 2 sleeping sa 0 master NETWO 3 sleeping sa 0 master CHECK 4 sleeping karenp PC Client 1 pubs SELEC 5 runnable karenp PC Client 0 pubs SELEC (5 rows affected) The spid column contains process identification numbers used in the TRANSACT-SQL KILL statement. The blk column contains the process ID of a blocking process, if there is one. A blocking process (which may have an exclusive lock) is one that is holding resources that another process needs. In the earlier example, process 4 (a SELECT on a table) is blocked by process 1 (a BEGIN TRANSACTION followed by an INSERT on the same table). Server processes cannot be killed. If you try to kill a server process, it is not killed and is still listed when you execute the sp_who system procedure. Sleeping processes that are killed don't die until they wake up, so you may still see them listed after you execute the KILL statement. You can get a report on the current locks and the spids of the processes holding them with the sp_lock system procedure. ╓┌─┌──────┌─────────────┌────────────┌──────┌────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── spid locktype table_id page dbname ----- ------------ ----------- ----- -------- 1 Sh_intent 16003088 0 master ──────────────────────────────────────────────────────────────────────────── 1 Sh_intent 16003088 0 master 4 Ex_extent 0 440 pubs 4 Ex_extent 0 504 pubs 4 Sh_table 112003430 0 pubs 4 Ex_table 240003886 0 pubs Permissions KILL permission defaults to the System Administrator and is not transferable. See Also sp_lock, sp_who LOAD DATABASE ──────────────────────────────────────────────────────────────────────────── Function Loads a backup copy of a user database and its transaction log that was created with DUMP DATABASE. Syntax LOAD DATABASE database_name FROM dump_device Examples load database pubs from diskdump1 The pubs database has been reloaded from the diskdump1 device. Options database_name The name of the database that has been created to receive data from a dumped backup copy. It can be either a newly created database with no data or an existing database. Loading dumped data to an existing database replaces existing data with the loaded data. The database into which you load a dumped database must be large enough to hold the dumped data. dump_device The logical name of the dump device from which you are loading the database. Logical dump device names are available in master..sysdevices.name or with the sp_helpdevice system procedure. Comments The specified database must not be in use during the load. Any data in the specified database is replaced by the loaded data. When the LOAD DATABASE statement is executed for loading from a diskette, the console program prompts the operator and waits for responses. If the console program is not running or if the operator does not answer its questions, the load cannot proceed. For more information, see Chapter 3, "Utility Programs." The recipient database must be large enough for dumped data: it must be large enough to hold the highest logical page that was used by the dumped database. If the recipient database is too small, SQL Server displays an error message that gives the required size. To recover a database, load the most recent database dump and load all the transaction log dumps made since that database dump. The database into which you are loading the data is replaced by this reconstruction. Loading a master database that has been damaged is done with a different procedure. See "DISK REFIT," "DISK REINIT," or the SQL Server System Administator's Guide for details. For more information on SQL Server's backup and recovery facilities, see the SQL Server System Administrator's Guide. Permissions LOAD DATABASE permission defaults to the Database Owner and is not transferable. See Also DBCC, DUMP DATABASE, DUMP TRANSACTION, LOAD TRANSACTION, sp_helpdevice LOAD TRANSACTION ──────────────────────────────────────────────────────────────────────────── Function Loads a backup copy of the transaction log. The DUMP TRANSACTION and LOAD TRANSACTION statements should be used only when the transaction log is stored on a different dump device than the rest of the database. Syntax LOAD TRANsaction database_name FROM dump_device Examples load transaction pubs from diskdump1 The transaction log for the pubs database, previously dumped to diskdump1, has been reloaded. Options database_name The name of the destination database. dump_device The logical name of the dump device from which the transaction log is being loaded. Logical dump device names are available in master..sysdevices.name or with the sp_helpdevice system procedure. Comments LOAD TRANSACTION loads a backup copy of the transaction log made with DUMP TRANSACTION. To recover a database, load the most recent database dump and then load (in order) all the transaction log dumps made since the database dump. The database is restored to its state at the time of the last transaction log dump. When the LOAD TRANSACTION statement is executed for loading from a diskette, the console program prompts the operator and waits for a response. If the console program is not running or if the operator does not answer its questions, the load cannot proceed. For more information, see Chapter 3, "Utility Programs." SQL Server checks timestamps on the dumped transaction log to make sure that the transactions are being loaded into the correct database and in the correct sequence. Do not attempt to modify syslogs with a DELETE, UPDATE, or INSERT statement. Doing so makes it impossible for SQL Server to recover correctly in case of a system failure. In addition, an attempt to delete all rows from syslogs causes SQL Server to get into an infinite loop that eventually fills up the entire database. For more information on SQL Server's backup and recovery facilities, see the SQL Server System Administrator's Guide. Permissions LOAD DATABASE permission defaults to the Database Owner and is not transferable. See Also DUMP DATABASE, DUMP TRANSACTION, LOAD DATABASE, sp_helpdevice Mathematical Functions ──────────────────────────────────────────────────────────────────────────── Function Return values commonly needed for operations on mathematical data. Mathematical function names are not keywords. Syntax Mathematical functions have the following syntax: function_name(parameters) Function names, parameters, and results are listed in the following table: ╓┌─────────┌────────────────────────────────┌────────────────────────────────╖ Function Parameters Result ──────────────────────────────────────────────────────────────────────────── ABS (numeric_expr) An absolute value of a given expression. The expression can be of integer, float, or money type. Results are of the same type as the numeric expression. ACOS (float_expr) An angle (in radians) whose cosine is the specified floating-point value. Function Parameters Result ──────────────────────────────────────────────────────────────────────────── ASIN (float_expr) An angle (in radians) whose sine is the specified floating-point value. ATAN (float_expr) An angle (in radians) whose tangent is the specified floating-point value. ATN2 (float_expr1, float_expr2) An angle (in radians) whose tangent is ( float_expr1/float_expr2). CEILING (numeric_expr) The smallest integer greater than or equal to the specified value. The expression can be of integer, float, or money type. Results are of the same type as the numeric expression. Function Parameters Result ──────────────────────────────────────────────────────────────────────────── the numeric expression. COS (float_expr) The trigonometric cosine of the specified angle (in radians). COT (float_expr) The trigonometric cotangent of the specified angle (in radians). DEGREES (numeric_expr) Degrees converted from radians. The expression can be of integer, float, or money type. Results are of the same type as the numeric expression. EXP (float_expr) The exponential value of the specified value. FLOOR (numeric_expr) The largest integer less than Function Parameters Result ──────────────────────────────────────────────────────────────────────────── FLOOR (numeric_expr) The largest integer less than or equal to the specified value. The expression can be of integer, float, or money type. Results are of the same type as the numeric expression. LOG (float_expr) The natural logarithm of the specified value. LOG10 (float_expr) The base 10 logarithm of the specified value. PI ( ) The constant value of 3.1415926535897936. POWER (numeric_expr, y) The value of numeric_expr to the power of y. The expression and y can be of integer, float, Function Parameters Result ──────────────────────────────────────────────────────────────────────────── and y can be of integer, float, or money type. Results are of the same type as numeric_expr. RADIANS (numeric_expr) Radians converted from degrees. The expression can be of integer, float, or money type. Results are of the same type as numeric_expr. RAND ([int_expr]) A random float number using int_expr as the optional seed. ROUND (numeric_expr, int_expr) A numeric expression rounded off to the precision specified in int_expr. The expression can be of integer, float, or money type. Results are of the same type as numeric_expr. Function Parameters Result ──────────────────────────────────────────────────────────────────────────── type as numeric_expr. SIGN (numeric_expr) Positive (+1), zero (0), or negative (-1). The expression can be of integer, float, or money type. Results are of the same type as numeric_expr. SIN (float_expr) The trigonometric sine of the specified angle (measured in radians). SQRT (float_expr) The square root of the specified value. TAN (float_expr) The trigonometric tangent of the specified angle (measured in radians). Function Parameters Result ──────────────────────────────────────────────────────────────────────────── ──────────────────────────────────────────────────────────────────────────── Examples Statement Result A. select floor(123.45) 123.000000 select floor(-123.45) -124.000000 select floor($123.45) $123.00 B. select ceiling(123.45) 124.000000 select ceiling(-123.45) -123.000000 select ceiling($123.45) $124.00 C. select round(123.4545, 2) 123.450000 select round(123.45, -2) 100.000000 The ROUND function always returns a value even if the length is illegal. If the specified length is positive and longer than the digits after the decimal point, 0 is added after the fraction digits. If the length is negative and larger than or equal to the digits before the decimal point, ROUND returns 0.00. The last digit in a floating-point number is always an estimate: round(123.9995, 3) = 123.999 while round(123.9996, 3) = 124.000 Comments Error traps are provided to handle domain or range errors of these functions. A user can set the options ARITHABORT or ARITHIGNORE, which respectively abort the query or return NULL when a domain error occurs. No warning message is displayed. If neither of these options is set, the system returns NULL and prints a warning message after the query is executed. See Also Conversion Function, Functions, SET, String Functions, Text/Image Functions Null Values ──────────────────────────────────────────────────────────────────────────── Function Mark columns having an unknown value (as opposed to those that have 0 or a blank as a value). NULL allows you to distinguish between a deliberate entry of zero (for numerical columns) or a blank (for character columns) and a nonentry (NULL for both numerical and character columns). Syntax In CREATE TABLE statements: column_name datatype [NULL | NOT NULL] In SELECT statements: WHERE column_name IS [NOT] NULL In INSERT statements: VALUES({constant | NULL} [, {constant | NULL}]...) In UPDATE statements: SET column_name = {expression | NULL} [, column_name = {expression | NULL}...] The ISNULL built-in function: isnull(expression, value) Examples A. create table test (t1 char(10) null, t2 char(10) not null) insert test values (null, "stuff") insert test (t2) values ("stuff") Because column t1 of table test is defined as accepting NULL in the CREATE TABLE statement (example A), these two INSERT statements are equivalent to each other. The user can explicitly insert NULL or allow the system to do it. NULL is inserted in the t1 column whenever the user inserts rows but does not specify a value for the column. B. select isnull(t1, "unknown") from test Example B selects all the rows from test and displays all the null values in column t1 with the value unknown. C. update titles set advance = null where title_id = "TC3218" D. select title_id, advance from titles where advance < $5000 or advance is null Options NULL Indicates that the user (or a program) has made no entry. The value is unknown rather than blank or 0. In CREATE TABLE statements, NULL after a column name means that if there is no default for this column, SQL Server assigns a null value whenever the user does not make an entry for this column at insert time. In SELECT statements, use IS [NOT] NULL to retrieve null values (allowed only if the column has been defined to allow null values in the CREATE TABLE statement). An expression with a bitwise or arithmetic operator evaluates to NULL if any of the operands is null. NOT NULL Is the default in CREATE TABLE statements. It can also be used in SELECT statements to retrieve all values except those that are NULL. Comments Only columns for which NULL was specified in the CREATE TABLE statement, and into which you have explicitly entered NULL or into which no data has been entered contain null values. (Avoid entering the character string "NULL" as data for a character column. Use "N/A" or "none" instead. When you wish to enter the value NULL explicitly, do not use single or double quotation marks.) NOT NULL (the default) in a CREATE TABLE statement means that NULL is not a legal value for this column. If there is no default for this column, an error message will be produced whenever the user does not make an entry for this column at insert time. In addition, the user cannot assign NULL as a value with an INSERT or UPDATE statement. Columns added to an existing table with the ALTER TABLE statement (except bit columns, which cannot be added to an existing table) must allow NULL because the initial values of the new columns in existing rows are set to NULL. When you create NULL columns with certain datatypes, SQL Server automatically converts them to a different internal datatype to allow the storage of null values. SQL Server does not inform the user of the type change, and the user should be concerned about it only if querying the system tables. The char datatype is automatically converted to varchar; binary to varbinary; datetime to datetimn; float to floatn; int, smallint, and tinyint to intn; and money to moneyn. An expression with an arithmetic or bitwise operator evaluates to NULL if any of the operands is NULL. For example, 1 + column1 evaluates to NULL if column1 is NULL. An expression with a comparison operator evaluates to false if any of the operands is NULL. This means that null values will never match another value (not even another NULL) when used with a comparison operator. Use IS NULL to find null values in queries (when the columns being searched are defined as allowing null values). If you try to find null values in columns defined as NOT NULL, SQL Server generates an error message. The result of a subquery that returns no values is NULL. The result of a divide-by-zero is NULL. Aggregate functions ignore null values, except COUNT(*), which includes them. For example, in the calculation avg(advance), null values in the advance column are not counted, either in calculating the total amount or the number of values. When the GROUP BY clause is used, null values form their own group. With ORDER BY, null values come before all others. In a SELECT clause with the DISTINCT keyword, which selects unique rows only, null values are considered duplicates of each other. Only one NULL is selected, no matter how many are encountered. Null values are never joined, not even to other null values. If you have two tables identical in structure except that one has NULL fields and some null values, while the other has NOT NULL fields, this difference makes it impossible to insert the data from the NULL table into the NOT NULL table with SELECT. If a subquery returns NULL, the statement results will not be what you expect. If you specify NOT NULL when you create a column and do not create a default for it, an error message is produced whenever a user fails to make an entry in that column. The following table illustrates the relationship between the existence of a default and the definition of a column as NULL or NOT NULL. The entries in the table show the result: No entry ──────────────────────────────────────────────────────────────────────────── NULL null default null null NOT NULL error default error error ──────────────────────────────────────────────────────────────────────────── For information on NULL as the default parameter for a procedure, see "CREATE PROCEDURE." The sp_help system procedure reports the nulltype of a column (whether or not the column accepts null values). See Also Aggregate Functions, CREATE PROCEDURE, Expressions, GROUP BY and HAVING Clauses, INSERT, Search Conditions, SELECT, Subqueries, UPDATE, sp_help ORDER BY Clause ──────────────────────────────────────────────────────────────────────────── Function Returns query results in the specified column(s) in sorted order. Syntax [ORDER BY {[table_name. | view_name.] column_name | select_list_number | expression} [ASC | DESC] [,{[table_name. | view_name.] column_name | select_list_number | expression} [ASC | DESC]]...] Examples A. select title, type, price from titles where price > $9.99 order by title B. select type, price, advance from titles order by type desc compute avg(price), avg(advance) by type Options ORDER BY Sorts the results by columns. You can sort up to 16 columns. In TRANSACT-SQL, you can use this statement for items that do not appear in the select list. You can sort by a column heading, a column name, an expression, or a number representing the position of the item in the select list (the select_list_number). ASC Indicates that results are to be sorted in ascending order. ASC is the default; if you do not specify ASC or DESC, ASC is assumed. DESC Indicates that results are to be sorted in descending order. Comments ORDER BY cannot be specified on text or image datatype columns. With ORDER BY, null values come before all others. If you use COMPUTE BY, you must also use an ORDER BY clause. The columns listed after COMPUTE BY must be identical to or a subset of those listed after ORDER BY, and must be in the same left-to-right order, start with the same expression, and not skip any expressions. For example, say the ORDER BY clause is order by a, b, c In that case, the COMPUTE BY clause can be any (or all) of these: compute by a, b, c compute by a, b compute by a You can use a column name, an expression, or a column heading (or alias) in the ORDER BY clause. The COMPUTE keyword can be used without BY to generate grand totals, grand counts, and so on. ORDER BY is optional if you use the COMPUTE keyword without BY. Subqueries cannot include an ORDER BY clause, a COMPUTE clause, or the INTO keyword. View definitions cannot include an ORDER BY clause, a COMPUTE CLAUSE, or the INTO keyword. The ORDER BY clause cannot be specified on columns containing text or image datatypes. See Also COMPUTE Clause, Expressions, GROUP BY and HAVING Clauses, SELECT, WHERE Clause Parameters ──────────────────────────────────────────────────────────────────────────── Function Values supplied to a stored procedure. Parameters are defined (optionally) when a stored procedure is created, and their values are supplied by the user when a procedure is executed. Syntax CREATE PROCEDURE [[database.]owner]procedure_name[;number] [[(]@parameter_name datatype [= default] [,@parameter_name datatype [= default]]...[)]] [WITH RECOMPILE] AS SQL_statements [EXECute] procedure_name [value [, value]...] or [EXECute] procedure_name [@parameter_name = value [,@parameter_name = value]...] Examples create procedure showtype @tabname varchar(18), @colname varchar(18) as select syscolumns.name, syscolumns.length, systypes.name from syscolumns, systypes, sysobjects where sysobjects.id = syscolumns.id and @tabname = sysobjects.name and @colname = syscolumns.name and syscolumns.type = systypes.type exec showtype sales, qty Options @parameter_name The name of the parameter or parameter to the stored procedure. Parameter names must be preceded by the "at" symbol (@) and conform to the rules for identifiers. datatype The datatype of the parameter. A length in parentheses after the datatype entry is required for some datatypes. See the "Datatypes" section for a list of SQL-Server-supplied datatypes and their syntax. default A default parameter value for the procedure. If a default is defined, a user can execute the procedure without giving a parameter. The default must be a constant. It can include the wildcard characters (%, _, [ ], and [^]) if the procedure uses the parameter name with the LIKE keyword. The default can be NULL. The procedure definition can specify that some action be taken if the parameter value is NULL. See "CREATE PROCEDURE" for examples. Comments Parameter names are optional in CREATE PROCEDURE statements; a procedure need not take any parameters. Parameter names are local to the procedure; that is, the same parameter names can be used in other procedures. Parameter names can take the place of constants only; they cannot be used in place of table names, column names, or the names of other database objects. However, the value of a parameter given at execution time can be an object name. See "EXECUTE" for examples. Enclose any parameter value that includes punctuation (such as an object name qualified by a database name or owner name) in single or double quotation marks. Parameters cannot be of text or image datatype. See Also CREATE PROCEDURE, DECLARE, Datatypes, DROP PROCEDURE, EXECUTE, Variables (Local and Global) PRINT ──────────────────────────────────────────────────────────────────────────── Function Prints a user-defined message on the user's screen. Syntax PRINT {"any ASCII text" | local variable | @version global variable} Examples A. if exists (select zip from authors where zip = '94705') print "Berkeley author" B. declare @msg char(50) select @msg = "What's up, doc?" print @msg What's up, doc? C. declare @tabname varchar(30) select @tabname = "titles" declare @username varchar(30) select @username = "ezekiel" declare @message varchar(255) select @message = "The table '" + @tabname + "' is not owned by the user '" + @username + ".'" print @message The table 'titles' is not owned by the user 'ezekiel.' Comments The message can be up to 255 characters long. The local variable must be of type char or varchar and must be declared within the batch or procedure in which it is used. The global variable must be of type char or varchar or automatically convertible to these types, such as @VERSION. You can generate a "formatted" message with more than one parameter by using the concatenation operator, as shown in example C. See Also DECLARE, RAISERROR, Variables (Local and Global) RAISERROR ──────────────────────────────────────────────────────────────────────────── Function Prints a user-defined error message on the user's screen and sets a system flag to record that an error has occurred. Syntax RAISERROR number {"text of message" | local variable} Examples create procedure showtable_sp @tabname varchar(18) as if @tabname is null begin raiserror 99999 "You must give a table name" end else begin select sysobjects.name, type, crdate, indid from sysindexes, sysobjects where sysobjects.name = @tabname and sysobjects.id = sysindexes.id end Comments The error message can be up to 255 characters long. Error numbers for user-defined error messages must be greater than 20,000. When RAISERROR is executed, the error number is placed in the global variable @ERROR, which stores the error number most recently generated by the system. The local variable must be of type char or varchar, and must be declared within the batch or procedure in which it is used. Use RAISERROR instead of PRINT when you want an error number stored in @ERROR. See Also DECLARE, PRINT, Variables (Local and Global) READTEXT ──────────────────────────────────────────────────────────────────────────── Function Reads text and image values, starting from a specified offset and reading a specified number of bytes. Syntax READTEXT [[database.]owner.]table_name.column_name text_ptr offset size [HOLDLOCK] Examples declare @val varbinary(30) select @val = textptr(blurb) from texttest where title_id = "BU7832" readtext texttest.blurb @val 1 5 This example selects the second through sixth characters of the blurb column from the texttest table. Options table_name.column_name The table name and column name. The table name must be included, but the database name and owner name are optional. text_ptr A valid text pointer. offset The number of bytes to skip before starting to read the text or image data. size The number of bytes of data to read. If size = 0, 4K bytes of data are read. HOLDLOCK Causes the text value to be locked for reads until the end of the transaction. Other users can read the value, but they cannot modify it. Comments Since the TEXTPTR function returns a long binary string, it is best to declare a local variable to hold the text pointer and then use the variable with READTEXT. The HOLDLOCK option causes the text value to be locked for reads until the end of the transaction. Other users can read the value, but they cannot modify it. The TEXTPTR function returns a pointer to the text or image column in the specified row or to the text or image column in the last row returned by the query if more than one row is returned. The value in the global variable @TEXTSIZE, which is the limit on the number of bytes of data to be returned, supersedes the size specified for READTEXT if it is less than the specified size for READTEXT. Use SET TEXTSIZE to change this limit. Permissions Users must have select permission on the table to use READTEXT. See Also Text/Image Datatypes, Text/Image Functions, WRITETEXT RECONFIGURE ──────────────────────────────────────────────────────────────────────────── Function Part of the procedure that sets configuration options, which control various aspects of SQL Server's memory allocation and performance. The System Administrator uses RECONFIGURE. Syntax RECONFIGURE [WITH OVERRIDE] Examples reconfigure The dynamic values in the master..sysconfigures system table are put into effect immediately. The static values take effect the next time SQL Server is started. Options WITH OVERRIDE Instructs SQL Server to accept the values for system variables supplied by the user (with sp_configure). User-supplied values that are unacceptable to SQL Server cause the RECONFIGURE statement to abort. If the user supplies values that SQL Server cannot accept, an error message telling the user to re-execute sp_configure is displayed. The WITH OVERRIDE clause must always be used when you set the allow updates variable to 1. Comments To display a list of the configuration options with their current values and the range of permissible values, execute sp_configure without parameters. The report looks like this: ╓┌───────┌─────────────────────┌────────┌────────┌─────────────────────┌───── ───────────────────────────────────────────────────────────────────────────── name minimum maximum config_value run_val ----------------- ------- ------- ------------ ------- recovery interval 1 32767 4 3 allow updates 0 1 0 0 user connections * 5 35 0 25 memory 1000 14000 0 1700 open databases 5 100 0 10 ───────────────────────────────────────────────────────────────────────────── open databases 5 100 0 10 locks 5000 50000 0 5000 open objects 100 10000 0 500 procedure cache 1 99 0 20 fill factor 0 100 0 0 time slice 50 1000 0 100 database size 2 10000 0 2 media retention 0 365 0 0 recovery flags 0 1 0 0 serial number 1 999999 0 0 ───────────────────────────────────────────────────────────────────────────── (14 rows affected) * The SQL Server maximum is 250; the effective number is based upon your system configuration and is stored in the global variable @MAX_CONNECTIONS. The config_value column of the report contains the value to which the configuration option was set with sp_configure (the value in sysconfigures.value). The run_value column contains the value SQL Server is using. This value changes for dynamic configuration options (recovery interval and allow updates) after you run the RECONFIGURE statement. This value changes for static configuration options after you shut down and restart SQL Server. This is the value in syscurconfigs.value. To change a configuration option, execute sp_configure with its parameters optname (the name of the configuration option) and optvalue (a legal value). Then execute the RECONFIGURE statement. (Include the WITH OVERRIDE clause to turn the allow updates variable on or to override a warning that the value you have specified is considered less than optimal by SQL Server.) If you have changed any of the static configuration options (all of them except allow updates and recovery interval), you must restart SQL Server before the new values take effect. SQL Server checks to make sure that the new values in sysconfigures are acceptable (SQL Server will run) and reasonable (SQL Server will run optimally). The values in sysconfigures.value and in the config_value column of the sp_configure report do not necessarily reflect the values currently being used by SQL Server for two reasons. First, the System Administrator may have updated sysconfigures.value but not executed the RECONFIGURE statement and/or not restarted SQL Server. Second, the default for most values in sysconfigures is 0, which indicates that SQL Server calculates optimum values for that parameter. Here are the contents of sysconfigures: ╓┌───────┌───────┌──────┌──────────────────────────────────────────────────── ───────────────────────────────────────────────────────────────────────────── config value comment ------ ----- ------------------------------------------------------ 101 4 Maximum recovery interval in minutes 102 0 Allow updates to system tables 103 0 Number of user connections allowed 104 0 Size of available physical memory in 2K pages ───────────────────────────────────────────────────────────────────────────── 105 0 Number of open databases allowed among all users 106 0 Number of locks for all users 107 0 Number of open database objects 108 0 Percentage of remaining memory used for procedure cache 109 0 Default fill factor percentage 110 0 Average time slice per process in milliseconds 111 0 Default database size in megabytes 112 0 Backup media retention period in days 113 0 Recovery flags ───────────────────────────────────────────────────────────────────────────── 114 0 Serial Number The value in the value column is the value that has been set with sp_configure. The value in the status column is 1 if the configuration option is dynamic and 0 if it is not. The overhead of each additional user connection (103) is 34K. The total size of available memory (104) must be changed when you add or remove memory boards. The total available physical memory minus the memory required for the operating system is available for SQL Server. The percentage of memory allocated to the procedure cache (108) and, by implication, to the buffer cache, can affect performance noticeably. You may have to change the database size (111) if the size of your model database increases substantially (for example, if you add a lot of your own stored procedures to model). You may want to change the default database size if most of your databases are larger than the value in syscurconfigs, which is 2 megabytes. For more information, see the SQL Server System Administrator's Guide. Permissions RECONFIGURE permission defaults to the System Administrator and is not transferable. See Also sp_configure RETURN ──────────────────────────────────────────────────────────────────────────── Function Exits unconditionally from a query or procedure. Statements following RETURN are not executed. Syntax RETURN Examples create procedure findrules @nm varchar(30) = null as if @nm is null begin print "You must give a user name" return end else begin select sysobjects.name, sysobjects.id, sysobjects.uid from sysobjects, master..syslogins where master..syslogins.name = @nm and sysobjects.uid = master..syslogins.suid and sysobjects.type = "R" end If no username is given as a parameter, the RETURN keyword causes the procedure to exit after a message has been sent to the user's screen. If a username is given, the names of the rules created by this user in the current database are retrieved from the appropriate system tables. Comments The RETURN keyword can be used at any point where you want to exit from a procedure. RETURN is immediate and complete: statements after RETURN are not executed. See Also BEGIN...END, IF...ELSE, WHILE REVOKE ──────────────────────────────────────────────────────────────────────────── Function Revokes permissions from users. Syntax Object permissions: REVOKE {ALL | permission_list} ON {table_name [(column_list)] | view_name [(column_list)] | stored_procedure_name} FROM {PUBLIC | name_list} Statement permissions: REVOKE {ALL | statement_list} FROM {PUBLIC | name_list} Examples A. use pubs /* Permissions can be revoked on objects ** in the current database only. */ revoke insert, delete on titles from mary, sales B. revoke update on titles (royalty, advance) from public C. revoke create database, create table from mary, john D. revoke all from mary E. revoke all on titles from mary Options ALL When used to revoke object permissions (first syntax format), this keyword specifies that all permissions applicable to the specified object are revoked. When used to revoke statement permissions (second syntax format), this keyword specifies that all statement permissions are revoked. (CREATE DATABASE permission can be granted only by the System Administrator.) permission_list A list of granted permissions. When permissions are revoked on a table or a view, the permission list can include one or more of the following: SELECT, INSERT, DELETE, and UPDATE. When permissions are revoked on columns, the permission list can include one or both of the following: SELECT and UPDATE. When permissions are revoked on stored procedures, the permission list can include EXECUTE only. If more than one permission is listed, separate them with commas. statement_list A list of granted statements. The statement list can include CREATE DATABASE (which can be granted only by System Administrator), CREATE DEFAULT, CREATE PROCEDURE, CREATE RULE, CREATE TABLE, CREATE VIEW, DUMP DATABASE, and DUMP TRANSACTION. If more than one statement is listed, separate them with commas. table_name The name of the table in the current database. Only one object can be listed for each REVOKE statement. column_list A list of columns, separated by commas, to which the permissions apply. If columns are specified, only SELECT and UPDATE permissions can be granted. view_name The name of the view in the current database. Only one object can be listed for each REVOKE statement. stored_procedure_name The name of the stored procedure in the current database. Only one object can be listed for each REVOKE statement. PUBLIC All users. name_list A list of usernames and/or groupnames, separated by commas. Comments Permissions can be granted or revoked on objects in the current database only. You can substitute the word "TO" for the word "FROM" in the REVOKE statement if you wish. User groups allow you to grant or revoke permissions to more than one user at a time. Each user can be a member of only one group. Groups are created with the sp_addgroup system procedure, which takes the groupname as its parameter. New users can be added to a group (other than public, the default group) with the sp_adduser system procedure. Users' groups can be changed with sp_changegroup. To remove a group, use the sp_dropgroup system procedure. Its parameter is the groupname. To display the members of a group, use the sp_helpgroup system procedure. Its optional parameter is the groupname. If you do not give a groupname, sp_helpgroup displays all the groups that exist and lists the members of each. GRANT and REVOKE statements are order sensitive. The statement that takes effect when there is a conflict is the most recently executed one. See Also GRANT, SETUSER, sp_addgroup, sp_adduser, sp_changedbowner, sp_changegroup, sp_dropgroup, sp_dropuser, sp_helpgroup, sp_helprotect, sp_helpuser ROLLBACK TRANSACTION ──────────────────────────────────────────────────────────────────────────── Function Rolls back a user-specified transaction to the last savepoint inside a transaction or to the beginning of a transaction. Syntax ROLLBACK TRANsaction [transaction_name | savepoint_name] Examples begin transaction royaltychange /* A user sets out to change the royalty split for the two authors of The Gourmet Microwave. Since the database would be inconsistent between the two updates, they must be grouped into a user-defined transaction. */ update titleauthor set royaltyper = 65 from titleauthor, titles where royaltyper = 75 and titleauthor.title_id = titles.title_id and title = "The Gourmet Microwave" update titleauthor set royaltyper = 35 from titleauthor, titles where royaltyper = 25 and titleauthor.title_id = titles.title_id and title = "The Gourmet Microwave" save transaction percentchanged /* After having updated the royaltyper entries for the two authors, the user inserts the savepoint percentchanged, and then experiments to see how a 10% increase in the book's price would affect the authors' royalty earnings. */ update titles set price = price * $1.1 where title = "The Gourmet Microwave" select (price * royalty * ytd_sales) * royaltyper from titles, titleauthor where title = "The Gourmet Microwave" and titles.title_id = titleauthor.title_id /* The transaction is rolled back to the savepoint with the ROLLBACK TRANSACTION statement.*/ rollback transaction percentchanged commit transaction /* End of royaltychange. */ Options transaction_name The name assigned to the transaction. It must conform to the rules for identifiers. savepoint_name The name assigned to the savepoint. It must conform to the rules for identifiers. Comments Use the form ROLLBACK TRANSACTION transaction_name. When you rollback a transaction, all of the transaction's statements or procedures are undone. It is as though you had never performed any of those statements on procedures. If no savepoint_name or transaction_name is given with the ROLLBACK TRANSACTION statement, the transaction is rolled back to the previous BEGIN TRANSACTION. A savepoint is a marker set by a user within a transaction. It is used to cancel part of a transaction with the statement ROLLBACK TRANSACTION savepoint_name. All of the statements or procedures between the savepoint and the ROLLBACK TRANSACTION are undone. After a transaction is rolled back to a savepoint, it must proceed to completion (with more SQL statements if desired and a COMMIT TRANSACTION statement), or it must be canceled altogether (by rolling it back to its beginning). There is no limit on the number of savepoints within a transaction. The ROLLBACK TRANSACTION statement must appear within a transaction; you can't roll back a transaction after COMMIT TRANSACTION has been entered. Define a transaction by enclosing SQL statements and/or stored procedures with BEGIN TRANSACTION and COMMIT TRANSACTION. Within a transaction, duplicate savepoint_names are allowed, but only the first instance of a savepoint_name is used. Permissions ROLLBACK TRANSACTION permission defaults to public; no permission is required to use it. See Also BEGIN TRANSACTION, COMMIT TRANSACTION, SAVE TRANSACTION Row Aggregate Functions ──────────────────────────────────────────────────────────────────────────── Function Generate summary values that appear as additional rows in the query results (unlike the aggregate function results, which appear as new columns). They allow you to see detail and summary rows in one set of results. Row aggregate functions (SUM, AVG, MIN, MAX, and COUNT) are used in a SELECT statement with the COMPUTE clause. You can calculate summary values for subgroups, and you can calculate more than one aggregate function for the same group. COMPUTE and the row aggregate functions are SQL Server enhancements to standard SQL. Syntax COMPUTE row_aggregate(column_name) [, row_aggregate(column_name)...] [BY column_name [, column_name]...] Examples select type, price from titles where price > $10 and type like "%cook" order by type, price compute sum(price) by type ╓┌─┌──────────────────┌──────────────────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── type price ----------- ---------- mod_cook 19.99 sum ---------- 19.99 type price ------------ ---------- ──────────────────────────────────────────────────────────────────────────── ------------ ---------- trad_cook 11.95 trad_cook 14.99 trad_cook 20.95 sum ---------- 47.89 (6 rows affected) This example calculates the sum of the price of each type of cookbook over $10. (See "COMPUTE Clause" for more examples and information.) Options row_aggregate One of the following: ╓┌─────────────────────────────────┌─────────────────────────────────────────╖ Aggregate Function Result ──────────────────────────────────────────────────────────────────────────── SUM([DISTINCT] expression) Returns the total of the [distinct] values in the numeric column AVG([DISTINCT] expression) Returns the average of the [distinct] values in the numeric column COUNT(*) Returns the number of non-null values in the column MAX(expression) Returns the highest value in the column MIN(expression) Returns the lowest value in the column ──────────────────────────────────────────────────────────────────────────── column_name The name of a column. It must be enclosed in parentheses. Only numeric columns can be used with SUM and AVG. One COMPUTE clause can apply the same function to several columns. When using more than one function, use more than one COMPUTE clause. BY Indicates that row aggregate function values are to be calculated for subgroups. Whenever the value of the BY item changes, row aggregate function values are generated. If you use BY, you must use ORDER BY. Listing more than one item after BY breaks a group into subgroups and applies a function at each level of grouping. Comments The row aggregate functions make it possible to retrieve detail and summary rows with one command. The aggregate functions, on the other hand, ordinarily produce a single value for all the selected rows in the table or for each group, and these summary values are shown as new columns. The following examples illustrate the differences: A. select type, sum(price), sum(advance) from titles where type like "%cook" group by type ╓┌───┌─────────────────────────────────────────┌─────────────┌───────────────╖ ──────────────────────────────────────────────────────────────────────────── type ---------- ---------- ------------ mod_cook 22.98 15,000.00 trad_cook 47.89 19,000.00 (2 rows affected) B. select type, price, advance from titles where type like "%cook" order by type compute sum(price), sum(advance) by type type price advance ---------- ------------ ------------ mod_cook 2.99 15,000.00 mod_cook 19.99 0.00 sum sum ──────────────────────────────────────────────────────────────────────────── sum sum ------------ ------------ 22.98 15,000.00 type price advance ---------- ------------ ------------ trad_cook 11.95 4,000.00 trad_cook 14.99 8,000.00 trad_cook 20.95 7,000.00 sum sum ------------ ------------ 47.89 19,000.00 (7 rows affected) The columns in the COMPUTE clause must appear in the select list. You can't use SELECT INTO in the same statement as a COMPUTE clause because statements that include COMPUTE generate tables that include the summary results, which are not stored in the database. If you use COMPUTE BY, you must also use an ORDER BY clause. The columns listed after COMPUTE BY must be identical to or a subset of those listed after ORDER BY, and must be in the same left-to-right order, start with the same expression, and not skip any expressions. For example, say the ORDER BY clause is order by a, b, c the COMPUTE BY clause can be any or all of these: compute by a, b, c compute by a, b compute by a You must use a column name or an expression in the ORDER BY clause; you cannot sort by a column heading. The COMPUTE keyword can be used without BY to generate grand totals, grand counts, and so on. ORDER BY is optional if you use the COMPUTE keyword without BY. When you sum or average integer data, SQL Server treats the result as an int value, even if the datatype of the column is smallint or tinyint. ──────────────────────────────────────────────────────────────────────────── WARNING To avoid overflow errors in DB-LIBRARY programs, make all variable declarations for results of averages or sums type int. ──────────────────────────────────────────────────────────────────────────── In a SELECT statement with a COMPUTE clause, the order of columns in the select list overrides the order of the aggregate functions in the COMPUTE clause. For example: ╓┌─┌───────────────────────────────┌───────────┌─────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── select a, b, c from test compute sum(c), max(b), min(a) a b c --------- ---------- ---------- ──────────────────────────────────────────────────────────────────────────── --------- ---------- ---------- 1 2 3 3 2 1 sum ======= 4 max ======= 2 min ========= 1 (3 rows affected) DB-LIBRARY programmers must be aware of this to put the aggregate function results in the right place. See Also Aggregate Functions, COMPUTE Clause, Functions, GROUP BY and HAVING Clauses, SELECT SAVE TRANSACTION ──────────────────────────────────────────────────────────────────────────── Function Sets a savepoint within a transaction. Syntax SAVE TRANsaction savepoint_name Examples begin transaction royaltychange /* A user sets out to change the royalty split for the two authors of The Gourmet Microwave. Since the database would be inconsistent between the two updates, they must be grouped into a user-defined transaction. */ update titleauthor set royaltyper = 65 from titleauthor, titles where royaltyper = 75 and titleauthor.title_id = titles.title_id and title = "The Gourmet Microwave" update titleauthor set royaltyper = 35 from titleauthor, titles where royaltyper = 25 and titleauthor.title_id = titles.title_id and title = "The Gourmet Microwave" save transaction percentchanged /* After having updated the royaltyper entries for the two authors, the user inserts the savepoint percentchanged and then experiments to see how a 10% increase in the book's price would affect the authors' royalty earnings. */ update titles set price = price * $1.1 where title = "The Gourmet Microwave" select (price * royalty * ytd_sales) * royaltyper from titles, titleauthor where title = "The Gourmet Microwave" and titles.title_id = titleauthor.title_id /* The transaction is rolled back to the savepoint with the ROLLBACK TRANSACTION statement. */ rollback transaction percentchanged commit transaction /* End of royaltychange. */ Options savepoint_name The name assigned to the savepoint. It must conform to the rules for identifiers. Comments A savepoint is a marker set by a user within a transaction. It is used to cancel part of a transaction with the statement ROLLBACK TRANSACTION savepoint_name. All of the statements or procedures after the savepoint are undone. After a transaction is rolled back to a savepoint, it must proceed to completion (with more SQL statements if desired and a COMMIT TRANSACTION statement), or it must be canceled altogether (by rolling it back to its beginning). To cancel an entire transaction, use the form ROLLBACK TRANSACTION transaction_name. All of the transaction's statements or procedures are undone. If no savepoint_name or transaction_name is given with the ROLLBACK TRANSACTION statement, the transaction is rolled back to the previous BEGIN TRANSACTION. The ROLLBACK TRANSACTION statement must appear within a transaction; you can't roll back a transaction after COMMIT TRANSACTION has been entered. Define a transaction by enclosing SQL statements and/or stored procedures with BEGIN TRANSACTION and COMMIT TRANSACTION. Within a transaction, duplicate savepoint_names are allowed, but only the first instance of a savepoint_name is used. Permissions SAVE TRANSACTION permission defaults to public; no permissions are required to use it. See Also BEGIN TRANSACTION, COMMIT TRANSACTION, ROLLBACK TRANSACTION Search Conditions ──────────────────────────────────────────────────────────────────────────── Function Set the conditions in a WHERE or HAVING clause. (Joins and subqueries are specified in the search conditions; see the "Joins" and "Subqueries" sections for full details.) Syntax Search conditions immediately follow the keywords WHERE or HAVING in a SELECT, INSERT, UPDATE, or DELETE statement. (HAVING search conditions differ from WHERE search conditions only in that aggregate functions are not allowed in WHERE conditions.) If you use more than one of the search conditions in a single statement, connect the conditions with AND or OR. {WHERE|HAVING}[NOT]expression comparison_operator expression {WHERE|HAVING}[NOT]column_name [NOT] LIKE "match_string" {WHERE|HAVING}[NOT]expression [NOT] BETWEEN expression AND expression {WHERE|HAVING}[NOT]expression [NOT] IN ({value_list|subquery}) {WHERE|HAVING}[NOT]EXISTS (subquery) {WHERE|HAVING}[NOT]expression comparison_operator{ANY|ALL}(subquery) {WHERE|HAVING}[NOT]column_name join_operator column_name {WHERE|HAVING}[NOT]boolean_expression {WHERE|HAVING}[NOT]boolean_expression {AND|OR} boolean_expression {WHERE|HAVING}[NOT]column_name IS [NOT] NULL (The last clause is allowed only where column_name has been defined as allowing NULL in the CREATE TABLE statement.) Examples A. where advance * $2 > ytd_sales * price B. where phone not like '415%' Example B finds all rows in which the phone number does not begin with 415. C. where au_lname like "[CK]ars[eo]n" Example C finds the rows for authors named Carson, Carsen, Karsen, and Karson. D. where advance < $5000 or advance is null E. where (type = "business" or type = "psychology") and advance > $5500 F. where ytd_sales between 4095 and 12000 G. where state in ('CA', 'IN', 'MD') Examples D-G find the rows in which the state is one of the three in the list. Options expression A column name, a constant, a function, any combination of column names, constants, and functions connected by arithmetic or bitwise operators, or a subquery. The arithmetic operators are as follows: ╓┌───────────────────────┌───────────────────────────────────────────────────╖ Symbol Meaning ──────────────────────────────────────────────────────────────────────────── ++ Addition - Subtraction * Multiplication / Division % Modulo ──────────────────────────────────────────────────────────────────────────── Symbol Meaning ──────────────────────────────────────────────────────────────────────────── ──────────────────────────────────────────────────────────────────────────── Addition, subtraction, division, and multiplication can be used on int, smallint, tinyint, float, and money columns. The modulo operator cannot be used with money or float columns. A modulo is the integer remainder after a division involving two integers. For example, 21 % 9 = 3 because 21 divided by 9 equals 2 with a remainder of 3. The bitwise operators are as follows: Symbol Meaning ──────────────────────────────────────────────────────────────────────────── & Bitwise and (two operands) | Bitwise or (two operands) Bitwise exclusive or (two operands) ~ Bitwise not (one operand) ──────────────────────────────────────────────────────────────────────────── The bitwise operators can be used on int, smallint, and tinyint columns only. All the bitwise operators translate the integer parameters into binary representation before evaluating them. (See "Expressions" for more information.) comparison_operator One of the following: ╓┌────────────────────┌──────────────────────────────────────────────────────╖ Symbol Meaning ──────────────────────────────────────────────────────────────────────────── = Equal to > Greater than < Less than >= Greater than or equal to <= Less than or equal to != Not equal to !> Not greater than !< Not less than ──────────────────────────────────────────────────────────────────────────── In comparing char and varchar data, < means closer to the beginning of the alphabet, and > means closer to the end of the alphabet. Case and special character evaluations depend on the sorting sequence used by the operating system on the machine on which SQL Server is located. On ASCII machines, for example, lowercase letters are greater than uppercase letters, and uppercase letters are greater than numbers. In comparing dates, < means earlier and > means later. Trailing blanks are ignored for the purposes of comparison. "Dirk" is the same as "Dirk ". Put quotation marks around all character and datetime data used with a comparison operator (= "Bennet", > "94609"). column_name The name of a column used in the comparison. Qualify the column name with its table or view name if there is any ambiguity. NOT Negates any boolean expression and keywords, such as LIKE, NULL, BETWEEN, IN, and EXISTS. LIKE Indicates that the following character string (enclosed by single or double quotation marks) is a matching pattern. LIKE is available for char, varchar, and datetime columns (but not to search for seconds or milliseconds). You can use the LIKE keyword and wildcard characters with datetime data as well as with char and varchar. When you use LIKE with datetime values, SQL Server converts the dates to the standard datetime format and then to varchar. Since the standard storage format doesn't include seconds or milliseconds, you cannot search for seconds or milliseconds with LIKE and a pattern. It is a good idea to use LIKE when you search for datetime values, since datetime entries can contain a variety of date parts. For example, if you insert the value "9:20" into a column named arrival_time, the clause WHERE arrival_time = "9:20" would not find it because SQL Server converts the entry into "Jan 1, 1900 9:20AM". However, the clause WHERE arrival_time LIKE "%9:20%" would find it. match_string A string of characters and wildcards enclosed in quotation marks. The wildcards are as follows: Symbol Meaning ──────────────────────────────────────────────────────────────────────────── % Any string of 0 or more characters _ (underscore) Any single character [ ] Any single character within the specified range ([a-f]) or set ([abcdef]) [^] Any single character not within the specified range ([^a-f]) or set ([^abcdef]) ──────────────────────────────────────────────────────────────────────────── The wildcard and the string must be enclosed in single or double quotation marks. For complete information, see "Wildcard Characters." IS [NOT] NULL Keywords that search for null values (or all values except null values). It is allowed only if the column has been defined to allow null values in the CREATE TABLE statement. An expression with a bitwise or arithmetic operator evaluates to NULL if any of the operands is null. BETWEEN The range-start keyword. Use AND for the range-end value. A range of BETWEEN x and y, unlike a range of > x < y, is inclusive. IN Allows you to select values that match any one of a list of values. The comparator can be a constant or a column name, and the list can be a set of constants or, more commonly, a subquery. (See "Subqueries" for information on using IN with a subquery.) Enclose the list of values in parentheses. value_list A list of values. Put single or double quotation marks around char, varchar, and datetime values, and separate each value from the following one with a comma. (See example E, earlier in this section.) EXISTS Is used with a subquery to test for the existence of some result from the subquery. (See "Subqueries" for more information.) subquery A restricted SELECT statement (ORDER BY and COMPUTE clauses and the INTO keyword are not allowed) in the WHERE or HAVING clause of a SELECT, INSERT, DELETE, or UPDATE statement, or in a subquery or expression. (See "Subqueries" for more information.) ANY Is used with <, >, or = and a subquery. It returns results when any value retrieved in the subquery matches the value in the WHERE or HAVING clause of the outer statement. (See "Subqueries" for more information.) ALL Is used with < or > and a subquery. It returns results when all values retrieved in the subquery match the value in the WHERE or HAVING clause of the outer statement. (See "Subqueries" for more information.) join_operator A comparison operator, =* or *=. (See "Joins" for more information.) boolean_expression An expression that returns true or false. (See "Expressions" for a full definition.) AND Joins two conditions and returns results when both of the conditions are true. When more than one logical operator is used in a statement, AND operators are normally evaluated first. However, you can change the order of execution with parentheses. OR Joins two conditions and returns results when either of the conditions is true. When more than one logical operator is used in a statement, OR operators are normally evaluated after AND operators. However, you can change the order of execution with parentheses. Comments WHERE and HAVING search conditions are identical, except that aggregate functions are not permitted in WHERE clauses (HAVING avg(price) > $20 is legal; WHERE avg(price) > $20 is not legal). See "Aggregate Functions" for information on the use of aggregate functions. You can use a HAVING clause without a GROUP BY clause. If there are columns in the select list that neither have aggregate functions applied to them nor are included in the query's GROUP BY clause (illegal in standard SQL), the meanings of HAVING and WHERE are somewhat different. In this situation, a WHERE clause restricts the rows that are included in the calculation of the aggregate function but does not restrict the rows returned by the query. Conversely, a HAVING clause restricts the rows returned by the query but does not affect the calculation of the aggregate function. See "GROUP BY and HAVING Clauses" for examples. See Also Aggregate Functions, DELETE, EXECUTE, Expressions, INSERT, Joins, SELECT, Subqueries, UPDATE, Wildcard Characters, sp_helpjoins SELECT ──────────────────────────────────────────────────────────────────────────── Function Retrieves rows from the database. Syntax SELECT [ALL | DISTINCT] select_list [INTO [[database.]owner.]table_name] [FROM [[database.]owner.]{table_name | view_name} [HOLDLOCK] [,[[database.]owner.]{table_name | view_name} [HOLDLOCK]...] [WHERE search_conditions] [GROUP BY [ALL] aggregate_free_expression [, aggregate_free_expression...] [HAVING search_conditions] [ORDER BY {[[[database.]owner.]{table_name. | view_name.}] column_name | select_list_number | expression} [ASC | DESC] [,{[[[database.]owner.]{table_name | view_name.}] column_name | select_list_number | expression} [ASC | DESC]]...] [COMPUTE row_aggregate(column_name) [, row_aggregate(column_name)...] [BY column_name [, column_name]...]] [FOR BROWSE] The keywords in the SELECT statement, as in all other SQL statements, must be used in the order shown in the syntax statement. Examples A. select * from publishers B. select pub_id, pub_name, city, state from publishers C. select "The publisher's name is", Publisher = pub_name, pub_id from publishers D. select pub_id, total = sum (ytd_sales) from titles where advance < $10000 and ytd_sales is not null group by pub_id having count(*) >1 E. select type, price, advance from titles order by type desc compute avg(price), sum(advance) by type compute sum(price), sum(advance) F. select type, price, advance from titles compute sum(price), sum(advance) G. select * into coffeetabletitles from titles where price > $20 H. select * into newtitles from titles where price > $25 and price < $20 Options ALL Includes all rows in the results. This is the default. DISTINCT Includes only unique rows in the results. Null values are considered equal for the purposes of the DISTINCT keyword; only one NULL is selected no matter how many are encountered. select_list One or more of the following: ■ Asterisk (*), representing all columns in create-table order ■ A list of column names in the order in which you want to see them ■ A column heading to replace the default column heading (the column name) in the following form: column_heading = column_name or column_name column_heading ■ An expression (a column name, constant, function, or any combination of column names, constants, and functions connected by arithmetic or bitwise operators, or a subquery) ■ A built-in function or an aggregate function ■ Any combination of these items See "Expressions" for information on arithmetic and bitwise operators and a more detailed definition of expressions. See "Functions" for information on functions. See "Subqueries" for information on subqueries. INTO Creates a new table based on the columns specified in the select list and the rows chosen in the WHERE clause. The new table must not already exist. Its name must be unique in the database and must conform to the rules for identifiers. See example G. You can use SELECT INTO to create a duplicate table with no data by having a false condition in the WHERE clause. See example H. You cannot use SELECT INTO inside a user-defined transaction. SELECT INTO is a two-step operation. The first step creates the new table. The second step inserts the specified rows into the new table. If the second step fails for any reason (hardware failure, exceeding a system resource, and so on), the new table will exist with no rows in it. The select into/bulkcopy option must be on (by executing sp_dboption) to be able to select into a permanent table. You do not have to have the select into/bulkcopy option on to select into a temporary table, since the tempdb temporary database is never recovered. While the select into/bulkcopy option is on, you are not allowed to dump the transaction log because these operations are not logged and changes are therefore not recoverable from transaction logs. In this situation, executing the DUMP TRANSACTION statement produces an error message instructing you to use DUMP DATABASE instead. By default, the select into/bulkcopy option is off in newly created databases. To change the default, turn this option on in the model database. FROM Indicates that particular tables and views are used in the SELECT statement. It is always required except when the select list contains no column names (that is, constants and arithmetic expressions only): select 5 x, 2 y, "the product is", 5*2 Result table_name, view_name Tables and views used in the SELECT statement. If there is more than one table or view in the list, separate their names with commas. The order of the tables and views after the FROM keyword makes no difference in the results. As the syntax indicates, you can query tables in different databases in the same statement. Table names and view names can be given aliases, either for convenience or to distinguish the different roles that a table or view play in a self-join or subquery. Give the table or view name, then a space, then the alias name, like this: select * from publishers pu, authors au HOLDLOCK Restricts a shared lock on a specified table or view by holding it until the completion of a transaction (instead of releasing the shared lock as soon as the required table or data page is no longer needed, whether or not the transaction has been completed). The HOLDLOCK statement applies only to the table or view for which it is specified and only for the duration of the transaction defined by the statement in which it is used. search_conditions The conditions for the rows that are retrieved. A search condition can include column names; expressions; arithmetic operators; comparison operators; the keywords NOT, LIKE, IS NULL, AND, OR, BETWEEN, IN, EXISTS, ANY, or ALL; subqueries; or any combination of these items. There is no limit on the number of search conditions that can be included in an SQL statement. WHERE search conditions and HAVING search conditions are identical with one exception: WHERE clauses cannot contain aggregate functions and HAVING clauses can. See the sections "Search Conditions," "WHERE Clause," "Expressions," and "Subqueries" for more information. GROUP BY Keywords that find a value for each group. These values appear as new columns in the results, rather than as new rows. When GROUP BY is used with standard SQL, each item in the select list must either have a fixed value in every row in the group or be used with aggregate functions that produce a single value for each group. TRANSACT-SQL has no such restrictions on the items in the select list. Also, with TRANSACT-SQL, you can group by any expression (though not by a column alias), while with standard SQL you can only group by a column. You can use the following aggregate functions with GROUP BY (expression is almost always a column name): ╓┌─────────────────────────────────┌─────────────────────────────────────────╖ Aggregate Function Result ──────────────────────────────────────────────────────────────────────────── SUM([DISTINCT] expression) Returns the total of the [distinct] values in the numeric column AVG([DISTINCT] expression) Returns the average of the [distinct] values in the numeric column COUNT([DISTINCT] expression) Returns the number of [distinct] non-null values in the expression COUNT(*) Returns the number of selected rows MAX(expression) Returns the highest value in the expression MIN(expression) Returns the lowest value in the expression Aggregate Function Result ──────────────────────────────────────────────────────────────────────────── ──────────────────────────────────────────────────────────────────────────── (See "GROUP BY" and "Aggregate Functions" for more information.) A table can be grouped by any combination of columns─that is, groups can be nested within each other. Note that you cannot group by a column heading; you must use a column name, an expression, or a number representing the position of the item in the select list. ALL All groups in the results, even those that don't have any rows that meet the search conditions. (See "GROUP BY and HAVING Clauses" for an example.) aggregate_free_expression An expression that includes no aggregate functions. HAVING Sets conditions for the GROUP BY clause, similar to the way that WHERE sets conditions for the SELECT clause. There is no limit on the number of conditions that can be included. You can use a HAVING clause without a GROUP BY clause. If there are columns in the select list that neither have aggregate functions applied to them nor are included in the query's GROUP BY clause (illegal in standard SQL), the meanings of HAVING and WHERE are somewhat different. In this situation, a WHERE clause restricts the rows that are included in the calculation of the aggregate function but does not restrict the rows returned by the query. Conversely, a HAVING clause restricts the rows returned by the query but does not affect the calculation of the aggregate function. See "GROUP BY and HAVING Clauses" for examples. ORDER BY Sorts the results by columns. You can sort up to 16 columns. In TRANSACT-SQL, you can use ORDER BY for items that do not appear in the select list. You can sort by a column name, a column heading (or alias), an expression, or a number representing the position of the item in the select list (the select_list_number). If you sort by select list number, the columns to which the ORDER BY clause refers must be included in the select list, and the select list cannot be *. ASC Indicates that results are to be sorted in ascending order, which is the default. DESC Indicates that results are to be sorted in descending order. COMPUTE Used with row aggregate functions (SUM, AVG, MIN, MAX, and COUNT) to generate control break summary values. The summary values appear as additional rows in the query results, allowing you to see detail and summary rows with one statement. You cannot use a SELECT INTO clause with COMPUTE. If you use COMPUTE BY, you must also use an ORDER BY clause. The columns listed after COMPUTE BY must be identical to or a subset of those listed after ORDER BY, and must be in the same left-to-right order, start with the same expression, and not skip any expressions. For example, say the ORDER BY clause is order by a, b, c Then the COMPUTE BY clause can be any (or all) of these: compute by a, b, c compute by a, b compute by a The COMPUTE keyword can be used without BY to generate grand totals, grand counts, and so on. ORDER BY is optional if you use the COMPUTE keyword without BY. If more than one item is listed after COMPUTE BY, the items must be a subset of those listed after ORDER BY and must be in the same order. (See "COMPUTE Clause" for details and examples.) FOR BROWSE DB-LIBRARY only. Allows you to perform updates while viewing data. FOR BROWSE is used only in an application program using DB-LIBRARY. Comments The keywords in the SELECT statement, as in all other SQL statements, must be used in the order shown in the syntax statement. Although it is not shown in the syntax, the keyword ALL can be used after SELECT for compatibility with other implementations of SQL. ALL is the default. Used in this context, ALL is the opposite of DISTINCT: all retrieved rows are included in the results, whether or not some are duplicates. Column headings and table aliases must conform to the rules for identifiers. If you have two tables identical in structure except that one has NULL fields and the other has NOT NULL fields, this difference makes it impossible to insert the data from the NULL table into the NOT NULL table with a SELECT, whether or not there is any NULL data. The length of returned text defaults to the size of the text. To change the length of returned text, use the SET TEXTSIZE statement. The limit on the length of text data returned with a SELECT statement defaults to 32K, or to the size set with SET TEXTSIZE. Permissions SELECT permission defaults to the owner of the table or view, who can transfer it to other users. See Also Aggregate Functions, COMPUTE Clause, CREATE TRIGGER, DELETE, Expressions, Functions, GROUP BY and HAVING Clauses, INSERT, Row Aggregate Functions, Search Conditions, UPDATE, WHERE Clause, Wildcard Characters, sp_dboption SET ──────────────────────────────────────────────────────────────────────────── Function Sets SQL Server query-processing options for the duration of the user's work session, or inside a trigger or stored procedure. Syntax SET {{ ARITHABORT | ARITHIGNORE | BACKGROUND | NOCOUNT | NOEXEC | OFFSETS {keyword_list} | PARSEONLY | PROCID | SHOWPLAN | STATISTICS IO | STATISTICS TIME } {ON | OFF} | ROWCOUNT number | TEXTSIZE n} Examples A. set showplan on go set noexec on go select * from publishers go For each query, SQL Server returns a description of the processing plan but does not execute it (example A). B. set rowcount 10 For each query, SQL Server stops processing the query after it returns the first ten rows (example B). Options ARITHABORT Aborts a query when an overflow or divide-by-zero error occurs during query execution. ARITHIGNORE Returns NULL when an overflow or divide-by-zero error occurs during a query. No warning message is displayed. If neither ARITHABORT nor ARITHIGNORE is set, SQL Server returns NULL and prints a warning message after the query is executed. BACKGROUND Puts processes generated by subsequent statements in the background and breaks the connection to the workstation. Once BACKGROUND is on, you can't do anything else. Your session is over; you must interrupt it and log in again. If you execute a stored procedure that sets the background option, your session becomes disconnected from your workstation. The connection is not re-established even when the procedure finishes executing. NOCOUNT Turns off the message returned at the end of each statement that tells you how many rows were affected by the statement. The global variable @ROWCOUNT is updated even when NOCOUNT is on. NOEXEC Compiles each query but does not execute it. NOEXEC is often used with SHOWPLAN. Once NOEXEC is turned on, no subsequent statements will be executed (including other SET statements) until NOEXEC is turned off. OFFSETS DB-LIBRARY only. Returns the offset (position in relation to the beginning of the query) of specified keywords in SQL statements. The keyword list is a list, separated with commas, that can include any of the following SQL constructs: SELECT, FROM, ORDER, COMPUTE, TABLE, PROCEDURE, STATEMENT, PARAM, and EXECUTE. The OFFSETS option is used only in an application program using DB-LIBRARY. PARSEONLY Checks the syntax of each query and returns any error messages without generating a sequence tree, compiling, or executing the query. Do not use PARSEONLY in a stored procedure. PARSEONLY returns offsets if the OFFSETS option is on and there are no errors (however, OFFSETS is used only in an application program using DB-LIBRARY). PROCID Returns the ID number of the stored procedure to DB-LIBRARY (not to the user) before sending rows generated by that stored procedure. ROWCOUNT number Causes SQL Server to stop processing the query after the specified number of rows are returned. To turn this option off, so that all rows are returned, use "SET ROWCOUNT 0". SHOWPLAN Generates a description of the processing plan for the query and immediately processes it unless NOEXEC is set. Do not use SHOWPLAN in a stored procedure. STATISTICS IO Displays the number of scans, the number of logical reads (pages accessed), and the number of physical reads (disk accesses) for each table referenced in the statement; STATISTICS IO displays the number of pages written for each statement. STATISTICS TIME Displays the time it took to parse and compile each command and the time it took to execute each step of the command. Times are given in timeticks, the exact value of which is machine-dependent. TEXTSIZE n Specifies the size in bytes of text type data to be returned with a SELECT statement. If you specify a textsize of 0, 32K bytes of data are returned with a SELECT statement. The @TEXTSIZE global variable stores the current setting. Comments If you use the SET statement inside a trigger or stored procedure, the option reverts to its former setting after the trigger or procedure executes. Do not use the SHOWPLAN or PARSEONLY option within a stored procedure. The SET options can be divided into four categories: ■ PARSEONLY, NOEXEC, SHOWPLAN, BACKGROUND, ROWCOUNT, and NOCOUNT control the way a query is executed. It doesn't make sense to set both PARSEONLY and NOEXEC to ON. The default setting for ROWCOUNT is 0 (return all rows); the default for the others is OFF. ■ The STATISTICS options display performance statistics after each query. The default setting for these options is OFF. ■ ARITHABORT and ARITHIGNORE handle exceptional cases during query execution. You cannot set both ARITHABORT and ARITHIGNORE to ON. The default setting for these options is OFF. ■ OFFSETS and PROCID are used by application programs to interpret results from SQL Server. The default setting for these options is ON. Any options set with a SET statement take effect at the end of the batch. You can combine SET statements and queries in the same batch, but the SET options won't apply to the queries in that batch. In the first example, the SET statement is combined with the query in the same batch; therefore, the SET option doesn't apply to the queries in that batch. In the second example, the SET statement is executed before the batch; therefore, the SET option does apply to the queries in that batch. A. set showplan on select * from publishers go pub_id pub_name city state ------- ---------------------- -------- ------ 0736 New Age Books Boston MA 0877 Binnet & Hardley Washington DC 1389 Algodata Infosystems Berkeley CA (3 rows affected) B. set showplan on go select * from publishers go STEP 1 The type of query is SELECT FROM TABLE publishers Nested iteration Table Scan pub_id pub_name city state ------- ---------------------- ------------ ------ 0736 New Age Books Boston MA 0877 Binnet & Hardley Washington DC 1389 Algodata Infosystems Berkeley CA (3 rows affected) See Also Batch Queries, Stored Procedures, Triggers SETUSER ──────────────────────────────────────────────────────────────────────────── Function Impersonates another user. This statement is used by a Database Owner. Syntax SETUSER ["username"] Examples setuser "mary" go grant select on authors to joe setuser go The Database Owner temporarily adopts Mary's identity in the database to grant permissions on authors, a table owned by Mary. Comments When the Database Owner uses the SETUSER statement, SQL Server checks the permissions of the user being impersonated instead of the permissions of the Database Owner. When the System Administrator uses the SETUSER statement, he or she retains full System Administrator permissions. The user being impersonated must be listed in the sysusers table of the database. The SETUSER statement remains in effect until another SETUSER statement is given or until the current database is changed with the USE statement. The SETUSER statement with no username re-establishes the Database Owner's original identity. SETUSER is used by the Database Owner to adopt the identity of another user, to use another user's database object, to grant permissions, to create an object, or for some other reason. The System Administrator can use SETUSER to create objects that will be owned by another user. However, since the System Administrator operates outside the permissions system, she or he cannot use SETUSER to acquire another user's permissions. No matter what SETUSER statements have been executed, the System Administrator always has permission to do everything. Permissions SETUSER permission defaults to the Database Owner and is not transferable. See Also Batch Queries, GRANT, REVOKE SHUTDOWN ──────────────────────────────────────────────────────────────────────────── Function Brings the system to a halt. This statement can only be executed by the System Administrator. Syntax SHUTDOWN [WITH NOWAIT] Options WITH NOWAIT Keywords that shut SQL Server down immediately, without performing checkpoints in every database. SQL Server exits after attempting to terminate all user processes and roll back any active transactions. Comments Unless the System Administrator uses the NOWAIT option, SHUTDOWN waits for currently executing SQL statements or stored procedures to finish and then performs a checkpoint in every database. SHUTDOWN minimizes the amount of work that automatic recovery has to do when the System Administrator restarts SQL Server. Permissions SHUTDOWN permission defaults to the System Administrator and cannot be transferred. String Functions ──────────────────────────────────────────────────────────────────────────── Function Perform various operations on binary data, character strings, or expressions, including concatenation. Syntax Built-in string functions return values commonly needed for operations on character data. String function names are not keywords. String functions have the following syntax: function_name(parameters) You can concatenate binary or character expressions like this: (expression + expression [+ expression} . . .) Function names, parameters, and results are as follows: ╓┌───────────┌───────────────────────────────┌───────────────────────────────╖ Function Parameters Result ──────────────────────────────────────────────────────────────────────────── ASCII (char_expr) The ASCII code value of the leftmost character of a character expression. CHAR (int_expr) A character converted from an ASCII code. The ASCII code should be between 0 and 255; otherwise, NULL is returned. Function Parameters Result ──────────────────────────────────────────────────────────────────────────── otherwise, NULL is returned. CHARINDEX ("Pattern", expression) The starting position of the specified pattern. The first parameter is the pattern. The second parameter is an expression, usually a column name, in which SQL Server searches for the pattern. Not implemented for text and image data. DATALENGTH (char_expr) The length of any type of data value or text field. A NULL string returns a length of 1. DIFFERENCE (char_expr1, char_expr2) The difference between the values of two character expressions as returned by the Function Parameters Result ──────────────────────────────────────────────────────────────────────────── expressions as returned by the SOUNDEX function. LOWER (char_expr) Lowercase converted from uppercase. LTRIM (char_expr) Data without leading blanks. REPLICATE (char_expr, int_expr) A character expression repeated a specified number of times. If int_expr is negative, a NULL string is returned. RIGHT (char_expr, int_expr) Part of a character string starting int_expr characters from the right. If int_expr is negative, a NULL string is returned. Function Parameters Result ──────────────────────────────────────────────────────────────────────────── RTRIM (char_expr) Data without trailing blanks. SOUNDEX (char_expr) A four-digit (SOUNDEX) code for use in evaluating the similarity of two strings. SPACE (int_expr) A string of repeated spaces. The number of spaces is equal to int_expr. If int_expr is negative, a NULL string is returned. STR (float_expr [, length [, Character data converted from decimal]]) numeric data. The length is the total length, including decimal point, sign, digits, and spaces. The decimal value is the number of spaces to the Function Parameters Result ──────────────────────────────────────────────────────────────────────────── is the number of spaces to the right of the decimal point. STUFF (char_expr1, start, length, Data where length characters char_expr2) have been deleted from char_expr1 at start and where char_expr2 has been inserted into char_expr1 at start. Not implemented for text data. SUBSTRING (expression, start, length) Part of a character or binary string. The first parameter can be a character or binary string, a column name, or an expression that includes a column name. (Do not use expressions that include functions or subqueries.) The second parameter specifies the Function Parameters Result ──────────────────────────────────────────────────────────────────────────── second parameter specifies the position at which the substring begins. The third parameter specifies the number of characters in the substring. Not implemented for text data. UPPER (char_expr) Uppercase converted from lowercase. ++ The + option indicates two or more concatenated character strings, binary strings, column names, or a combination of them. Enclose character strings in single or double quotation marks. ──────────────────────────────────────────────────────────────────────────── Function Parameters Result ──────────────────────────────────────────────────────────────────────────── ──────────────────────────────────────────────────────────────────────────── Examples A. select Name = (au_lname + ", " + au_fname) from authors Example A displays author names under the column heading Name in last-name, first-name order with a comma after the last name, that is, Bennet, Abraham. B. select au_lname, substring(au_fname, 1, 1) from authors Example B displays the last name and first initial of each author, that is, Bennet A. C. select (au_lname + ", " + substring(au_fname, 1, 1) + ".") from authors Example C displays the last name and first initial of each author, that is, Bennet, A. D. select substring ((pub_id + title_id), 1, 6) from titles Example D displays the first six characters of the concatenation of each pub_id and title_id. In other words, it displays all four characters of the pub_id and the first two characters of the title_id. E. select charindex("wonderful", notes) from titles where title_id = "TC3218" Example E returns the position at which the pattern "wonderful" begins in the notes column of titles. Statement Result F. select right("abcde", 3) cde select right("abcde", 6) abcde G. select str(123.45, 6, 2) 123.45 H. select stuff("abc", 2, 1, "xyz") axyzc Comments All string functions and concatenation can be used on char, varchar, binary, and varbinary datatypes. In addition, the REPLICATE and RIGHT string functions can be used on datetime datatypes. String functions can be nested. See examples C and D. When you use constants with string functions, enclose them in single or double quotation marks. String functions can be used in the select list, in the WHERE clause, and anywhere an expression is allowed. Length and decimal parameters to STR (if supplied) should be non-negative. The default length is 10. The number is rounded to an integer by default or if the decimal parameter is 0. The specified length should be at least equal to or greater than the part of the number before the decimal point plus the number's sign (if any). If float_expr exceeds the specified length, the string returns "**" of the specified length: select str(123.45, 2, 2) -- ** (1 row affected) Do not use a function or subquery as the float_expr in the STR function. A short float_expr is right justified in the specified length, and a long float_expr is truncated to the specified number of decimal places. The STUFF function inserts a string into another string. It deletes a specified length of characters from char_expr1 at the start position. It then inserts char_expr2 into char_expr1 at the start position. If the start position or the length is negative, a NULL string is returned. If the start position is longer than char_expr1, a NULL string is returned. If the length to delete is longer than char_expr1, it is deleted to the first character in char_expr1: select stuff("abc", 2, 3, "xyz") ── axyz (1 row affected) The SOUNDEX function converts an alpha string to a four-digit code for comparison use. All characters that are nonalpha will be used to terminate the compression. This function always returns some value. The DIFFERENCE function compares two strings and evaluates the similarity between them, returning a value from 0 to 4. Four is the best match. For example: ╓┌──────────────┌────────────────────────────────────────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── select difference("abc", "abc") --------- 4 (1 row affected) select difference("abc", "abd") --------- 3 (1 row affected) See Also Conversion Function, Functions, Mathematical Functions, SELECT, Text/Image Functions Subqueries ──────────────────────────────────────────────────────────────────────────── Function Nest a SELECT statement inside a SELECT, INSERT, UPDATE, or DELETE statement, another subquery, or anywhere an expression is allowed (if it returns a single value). A subquery is always enclosed in parentheses. Many SQL statements that include subqueries can be alternatively formulated as joins. (See "Joins" for more information on joining; see "SELECT" for more information on the full SELECT statement; see "WHERE Clause" for more information on search conditions and WHERE clause syntax.) Syntax A subquery can appear anywhere an expression can be used, as long as it returns a single value. It is usually nested inside the WHERE or HAVING clause of an outer SELECT, INSERT, UPDATE, DELETE, or another subquery. A subquery can be used in the following contexts: A. expression comparison_operator [ANY | ALL] (subquery) B. expression [NOT] IN (subquery) C. [NOT] EXISTS (subquery) A subquery has the following restricted SELECT syntax: ( SELECT [DISTINCT] subquery_select_list [FROM [[database.]owner.]{table_name | view_name} [HOLDLOCK] [, [[database.]owner.]{table_name | view_name} [HOLDLOCK]]...] [WHERE search_conditions] [GROUP BY aggregate_free_expression [, aggregate_free_expression]...] [HAVING search_conditions] ) Examples A. select distinct pub_name from publishers where pub_id in (select pub_id from titles where type = "business") B. select distinct pub_name from publishers where "business" in (select type from titles where pub_id = publishers.pub_id) C. select distinct pub_name from publishers where exists (select * from titles where pub_id = publishers.pub_id and type = "business") Examples A, B, and C find the names of publishers who have published business books. D. select au_lname, au_fname from authors where au_id in (select au_id from titleauthor where title_id in (select title_id from titles where type = "popular_comp")) Example D finds the names of authors who have participated in writing at least one popular computing book. E. select title from titles where advance > all (select advance from titles, publishers where titles.pub_id = publishers.pub_id and pub_name = "Algodata Infosystems") Example E finds the books that commanded an advance greater than the largest advance paid by Algodata Infosystems. F. select au_lname, au_fname from authors where city = any (select city from publishers) Example F finds the authors that live in the same city as some publisher. Options DISTINCT Cannot be used with subqueries that include a GROUP BY clause. subquery_select_list A restricted select list for use with subqueries introduced with IN or a comparison operator. It consists of one expression. If a column name is used in the WHERE clause of the outer statement, a column name in the subquery select list must be join compatible with it. The select list must consist of only one column name except for the EXISTS subquery, in which case the asterisk (*) is usually used in place of the single column name. Do not specify more than one column name. expression A column name; a constant; a function; any combination of column names, constants, and functions connected by arithmetic or bitwise operators; or a subquery. (The three formats for subqueries in the "Syntax" section are all considered expressions and can be substituted wherever the word "expression" appears as long as the subquery returns a single value.) The arithmetic operators are as follows: ╓┌───────────────────────┌───────────────────────────────────────────────────╖ Symbol Meaning ──────────────────────────────────────────────────────────────────────────── ++ Addition - Subtraction * Multiplication / Division % Modulo ──────────────────────────────────────────────────────────────────────────── Addition, subtraction, division, and multiplication can be used on int, smallint, tinyint, float, and money columns. The modulo operator can be used with all integer datatypes but cannot be used with money or float columns. A modulo is the integer remainder after dividing two integers. For example, 21 % 9 = 3 because 21 divided by 9 equals 2 with a remainder of 3. The bitwise operators are as follows: Symbol Meaning ──────────────────────────────────────────────────────────────────────────── & Bitwise and (two operands) | Bitwise or (two operands) Bitwise exclusive or (two operands) ~ Bitwise not (one operand) ──────────────────────────────────────────────────────────────────────────── The bitwise operators can be used on int, smallint, and tinyint columns only. All the bitwise operators translate the integer parameters into binary representation before evaluating them. (See "Expressions" for more information.) comparison_operator One of the following: ╓┌────────────────────┌──────────────────────────────────────────────────────╖ Symbol Meaning ──────────────────────────────────────────────────────────────────────────── = Equal to > Greater than < Less than >= Greater than or equal to <= Less than or equal to Symbol Meaning ──────────────────────────────────────────────────────────────────────────── <= Less than or equal to != Not equal to !> Not greater than !< Not less than ──────────────────────────────────────────────────────────────────────────── In comparing char and varchar data, < means closer to the beginning of the alphabet, and > means closer to the end of the alphabet. Case and special character evaluations depend on the collating sequence of the machine on which SQL Server is located. On ASCII machines, for example, lowercase letters are greater than uppercase letters, and uppercase letters are greater than numbers. In comparing dates, < means earlier and > means later. Trailing blanks are ignored for comparison purposes. So, for example, "Dirk" is the same as "Dirk ". Put quotation marks around all character and datetime data used with a comparison operator (= "Bennet", > "94609"). ANY Is used with >, <, or = and a subquery. > ANY is true (and the subquery returns results) when the value in the outer query is greater than any value in the list (that is, greater than the minimum value in the list). < ANY is true (and the subquery returns results) when the value in the outer query is less than any value in the list (that is, less than the maximum value in the list). ALL Is used with > or < and a subquery. > ALL means that the value being compared must be greater than all values in the list (that is, greater than the maximum value). < ALL means that the value being compared must be less than all values in the list (that is, less than the minimum value). IN Returns results when any value in the subquery matches the value in the WHERE clause of the outer statement. (IN is equivalent to = ANY.) IN can also be used with a list of values enclosed in parentheses. See "WHERE Clause" for details. subquery A query with the restricted SELECT syntax shown in the "Examples" section. It cannot include ORDER BY or COMPUTE clauses or the INTO keyword. If it is introduced by an unmodified comparison operator (and therefore returns a single value), it cannot include GROUP BY and HAVING clauses. The select list must consist of only one column name, except for the EXISTS subquery, in which case the asterisk (*) is almost always used. (SQL Server does not check the select list of a subquery introduced with EXISTS.) A subquery can contain another subquery. A subquery is legal in any TRANSACT-SQL syntax that allows an expression as long as it returns a single value. EXISTS Is used with a subquery to test for the existence of some result from the subquery. EXISTS is not preceded by an expression. The subquery that follows the EXISTS keyword is enclosed in parentheses. If the NOT keyword is used, the WHERE clause is satisfied if no rows are returned by the subquery. Since EXISTS tests only for the existence of a condition, the asterisk (*) is almost always used as the select list. Comments A subquery can be used anywhere an expression can be used if it returns a single value. Note that the syntax for EXISTS is somewhat different than the syntax for the other keywords; it does not take an expression between WHERE and itself. The two queries following, which are semantically equivalent, illustrate the difference: A. select distinct pub_name from publishers where exists (select * from titles where pub_id = publishers.pub_id and type = "business") B. select distinct pub_name from publishers where pub_id in (select pub_id from titles where type = "business") A correlated (or repeating) subquery depends on the outer query for its values. It is executed repeatedly, once for each row that might be selected by the outer query. Here's an example: select distinct au_lname, au_fname from authors where 100 in (select royaltyper from titleauthor where titleauthor.au_id = authors.au_id) The subquery in this statement cannot be evaluated independently of the outer query. It needs a value for authors.au_id, but this value changes as SQL Server examines different rows in authors. A correlated subquery can also be used in the HAVING clause of an outer query. This example finds the types of books for which the maximum advance is more than twice the average for the group: select t1.type from titles t1 group by t1.type having max(t1.advance) >=all (select 2 * avg(t2.advance) from titles t2 where t1.type = t1.type) See Also DELETE, EXECUTE, Expressions, Functions, GROUP BY and HAVING Clauses, INSERT, Joins, SELECT, UPDATE, WHERE Clause System Functions ──────────────────────────────────────────────────────────────────────────── Function Return special information from the database. Syntax function_name(parameters) Function names, parameters, and results are listed in the following table. Where the parameter is optional, the function returns the current value. ╓┌────────────┌──────────────────────────────┌───────────────────────────────╖ Function Parameters Result ──────────────────────────────────────────────────────────────────────────── COL_LENGTH ("object name", "column name") The length of column COL_NAME (object ID #, column ID #) The column name DB_ID (["database name"]) The database ID # DB_NAME ([database ID #]) The database name Function Parameters Result ──────────────────────────────────────────────────────────────────────────── DB_NAME ([database ID #]) The database name HOST_ID ( ) The workstation ID # HOST_NAME ( ) The workstation name INDEX_COL ("object name", index ID #, The indexed column name key #) ISNULL (expression, value) NULL entries with the specified value OBJECT_ID ("database object name") The database object ID # OBJECT_NAME (database object ID #) The database object name SUSER_ID (["server username"]) The user's login ID # SUSER_NAME ([server user ID #]) The user's login ID Function Parameters Result ──────────────────────────────────────────────────────────────────────────── SUSER_NAME ([server user ID #]) The user's login ID USER_ID (["username"]) The user's database ID number USER_NAME ([user ID number]) The username ──────────────────────────────────────────────────────────────────────────── Examples A. select user_id("harold") Example A returns harold's ID #, 13. B. select user_name(13) Example B returns the login ID for user #13, harold C. select avg(isnull(price, $10.00)) from titles Example C finds the average of the prices of all titles, substituting the value $10.00 for all NULL entries in price. D. select isnull(price, 0) from titles Example D finds all rows with null values in price and displays the null value as "0". Comments When the parameter to a system function is optional, the current database, host computer, server user, or database user is assumed. Built-in functions must always be followed by parentheses. The system functions can be used in the select list, in the WHERE clause, and anywhere an expression is allowed. The System Administrator's server user ID is always 1. Server user ID -1 is always reserved for a guest account. The user ID of the Database Owner is always 1. The USER_NAME function returns the name of the current user if the parameter is less than zero. See Also Functions, SELECT System Procedures ──────────────────────────────────────────────────────────────────────────── Function Stored procedures that SQL Server supplies to update and report from the system tables. Comments Chapter 2 of this manual contains a complete explanation of each system procedure. System procedures are created in the master database. They are owned by the master Database Owner. Since system procedures are located in the master database, their permissions are also set there. In other words, the System Administrator or master Database Owner must set up permissions for the system procedures in the master database, not from user databases. You can run system procedures from any database. If a system procedure is executed from a database other than master, any references to system tables are mapped to the database from which the procedure is being run. For example, if the Database Owner of pubs runs sp_adduser from pubs, the new user is added to pubs..sysusers. The names of all system procedures begin with "sp_". You can create your own system procedures that can be executed from any database. Simply create a stored procedure in master and give it a name that begins with "sp_". System procedures (for example, sp_help) do not work on temporary tables because a procedure cannot access a temporary object unless it created the object itself during the current session. Some of the system procedures can be run only by Database Owners. These procedures make sure that the user executing the procedure is the owner of the database from which they are being executed. Other system procedures can be executed by any user who has been granted EXECUTE permission on them, but this permission must be granted in master. Therefore, a user can either have permission to execute a system procedure in all databases or in none of them. Authorized SQL Server users (those who have a login ID) who are not listed in master..sysusers are treated as guest in master. The guest user inherits the permissions granted to public (in the setup program), including permission to execute certain system procedures (for example, all the sp_help procedures). To deny a user permission on a system procedure, the System Administrator or master Database Owner must add him or her to master..sysusers and write a REVOKE statement on that procedure. Thus, the owner of a user database cannot directly control permissions on the system procedures within his or her own database. When the parameter value for a system procedure contains punctuation marks or embedded blanks, it must be enclosed in single or double quotation marks. When the parameter is an object name and the object name is qualified by a database name or owner name, the entire name must be enclosed in single or double quotation marks. Several system procedure tables in the master database are used by the system procedures to convert internal system values (for example, status bits) into readable English. One of them, spt_values, is used by a wide variety of system procedures: sp_configure, sp_dboption, sp_depends, sp_help, sp_helpdb, sp_helpdevice, sp_helpindex, sp_helpkey, sp_helprotect, and sp_lock. The table spt_values is never updated. To see how it is used, execute sp_helptext to look at the text for one of the system procedures that references it. The other system procedure tables are spt_monitor and spt_committab. In addition, several of the system procedures create and then drop temporary tables: sp_helpdb creates #spdbdesc, sp_helpdevice creates #spdevtab, and sp_helpindex creates #spindtab. See Chapter 2, "System Procedures," for a complete description of the system procedures. See Also CREATE PROCEDURE, EXECUTE Text/Image Datatypes ──────────────────────────────────────────────────────────────────────────── Function The text datatype stores extremely long (up to 231- 1 or 2,147,483,647) byte strings of printable characters. The image datatype stores extremely long (up to 231- 1 or 2,147,483,647) bytes of hexadecimal-encoded binary data. Syntax The data values for a column of type text or image are stored and managed as a linked list of data pages, but they appear to the user as if they were stored in a table row. A text or image datatype definition does not include a length. column_name {text | image} Examples A. create table texttest (title_id varchar(6), blurb text, pub_id char(4)) B. insert texttest values ("BU7832", "Straight Talk About Computers is an annotated analysis of what computers can do for you: a no-hype guide for the critical user.", "1389") C. create table imagetest (image_id varchar(6), imagecol image, graphic_id char(4)) D. insert imagetest values ("94732", 0xFFFF1111222223333444445555566666FFFFF, "1389") E. select blurb from texttest where blurb like "Straight Talk%" Comments When you initialize a text or image column with an INSERT, UPDATE, or WRITETEXT, SQL Server assigns a text pointer and allocates an entire data page, 2K bytes, to hold the text or image value. Text values, like char and varchar values, require quotation marks in an INSERT statement. SELECT statements returning text and image values return all the data, up to the limit specified in the global variable @TEXTSIZE. The SET TEXTSIZE statement is used to change this value. The initial value of @TEXTSIZE is 32K; the maximum value for @TEXTSIZE is 231- 1 (that is, 2,147,483,647). You can explicitly convert text values to char or varchar with the CONVERT function, but you are limited to the maximum length of these datatypes, 255 bytes. If you do not specify a length, the converted value has a default length of 30 characters. Implicit conversion is not supported. Text and image values are stored on a linked list of data pages. A text or image column in a table contains a pointer to the first data page or value fragment for that entry, and thus to the entire text or image value. Image values, like binary and varbinary values, must be preceded by "0x" for an INSERT statement. You can explicitly convert image values to binary or varbinary with the CONVERT function, but you are limited to the maximum length of these datatypes, 255 bytes. If you do not specify a length, the converted value has a default length of 30 bytes. Implicit conversion is not supported. There are a number of restrictions on the use of text and image columns. They cannot be used under the following conditions: ■ As parameters to stored procedures. It is also illegal to pass text or image values to stored procedures. ■ As local variables. ■ In an ORDER BY clause. ■ In a COMPUTE clause. ■ In a GROUP BY clause. ■ In an index. ■ In a search clause (WHERE clause) except with the LIKE expression. ■ In a join clause. Use the LIKE expression to search for a text pattern. Use the PATINDEX function to search for the starting position of the first occurrence of a specified pattern in a text column. See Also Datatypes, Functions, READTEXT, Text/Image Functions, WRITETEXT Text/Image Functions ──────────────────────────────────────────────────────────────────────────── Function Return values commonly needed for operations on text and image data. Text and image built-in function names are not keywords. Syntax function_name(parameter) Function names, parameters, and results are listed in the following table. ╓┌─────────────┌──────────────────────────────┌──────────────────────────────╖ Function Parameters Result ──────────────────────────────────────────────────────────────────────────── PATINDEX ("pattern", column_name) The starting position of the first occurrence of pattern in the specified column. You can use wildcard characters in pattern. SET TEXTSIZE {n 0} The limit, in bytes, of the text or image data to be returned with a SELECT statement. The current setting is stored in the @TEXTSIZE global variable. n is an integer that specifies the limit on the number of bytes to be returned. 0 restores the default limit (32K bytes.) Function Parameters Result ──────────────────────────────────────────────────────────────────────────── (32K bytes.) TEXTPTR (column_name) The text pointer value in varbinary format. The text pointer is checked to ensure that it points to the first text page. TEXTVALID ("table_name.column_name", Checks if a given text text_ptr) pointer is valid and returns 1 if the pointer is valid and 0 if the pointer is invalid. Note that the identifier for the text column must include the table name. ──────────────────────────────────────────────────────────────────────────── Function Parameters Result ──────────────────────────────────────────────────────────────────────────── ──────────────────────────────────────────────────────────────────────────── Examples A. declare @val varbinary(30) select @val = textptr(blurb) from texttest where title_id = "BU7832" readtext texttest.blurb @val 1 5 Example A uses the TEXTPTR function to locate the text column, blurb, associated with title_id BU7832 in the texttest table. The text pointer is put into a local variable @val and supplied as a parameter to the READTEXT statement, which returns 5 bytes starting at the second byte (offset of 1). B. set textsize 100 C. select title_id, textptr(blurb) from texttest D. select textvalid ("texttest.blurb", textptr(blurb)) from texttest Comments "SELECT textptr(column_name) from my_table" returns a long binary string. It is a good idea to put this string into a local variable, as in the first example, and use it by reference. Explicit conversion is supported from text to char, and image to binary and varbinary, but you are limited to the maximum length for these datatypes, 255 bytes. If you do not specify a length, the default value of 30 bytes is used. Implicit conversion of text or image to another datatype is not supported. You cannot create an index on a text or image column. A text value cannot appear in a search clause or join clause, except in WHERE clauses using LIKE. A text value cannot be a local variable or a parameter to a stored procedure. It is also illegal to pass text values to stored procedures. ORDER BY cannot be specified on text type columns. See Also Conversion Function, Functions, Mathematical Functions, READTEXT, String Functions, Text/Image Datatypes TRUNCATE TABLE ──────────────────────────────────────────────────────────────────────────── Function This example removes all rows in a table as quickly as possible. Syntax TRUNCATE TABLE [[database.]owner.]table_name Examples truncate table authors This example removes all data in the authors table. Comments TRUNCATE TABLE restores the database to the same state it would be in if the CREATE TABLE and CREATE INDEX statements had been run on the table. The table, though empty of data, continues to exist until you give a DROP TABLE statement. TRUNCATE TABLE and DELETE without any conditions are functionally equivalent, but TRUNCATE TABLE is faster. TRUNCATE TABLE removes whole data pages and is not logged; DELETE removes rows one at a time and logs these transactions. Both DELETE and TRUNCATE TABLE reclaim the space occupied by the data and its associated indexes. A TRUNCATE TABLE statement is not caught by a DELETE trigger. Although a TRUNCATE TABLE statement is, in effect, like a DELETE without a WHERE clause (it removes all rows), it is not logged, and therefore cannot "fire" a trigger. Since permission for the TRUNCATE TABLE statement defaults to the table owner and is not transferable, only the table owner need worry about inadvertently circumventing a DELETE trigger with a TRUNCATE TABLE statement. TRUNCATE TABLE permission defaults to the table owner and is not transferable. See Also DELETE, DROP TABLE UPDATE ──────────────────────────────────────────────────────────────────────────── Function Changes data in existing rows, either by adding new data or modifying existing data. Syntax UPDATE [[database.]owner.]{table_name | view_name} SET [[[database.]owner.]{table_name. | view_name.}] column_name1 = {expression1 | NULL} [, column_name2 = {expression2 | NULL}...] [WHERE search_conditions] UPDATE [[database.]owner.]{table_name | view_name} SET [[[database.]owner.]{table_name. | view_name.}] column_name1 = {expression1 | NULL} [, column_name2 = {expression2 | NULL}...] [FROM [[database.]owner.]{table_name | view_name} [, [[database.]owner.]{table_name | view_name}]...] [WHERE search_conditions] Examples A. update authors set au_lname = "MacBadden" where au_lname = "McBadden" Example A changes all the McBaddens in the authors table to MacBaddens. B. update titles set ytd_sales = ytd_sales + qty from titles, sales where titles.title_id = sales.title_id and sales.date in (select max(sales.date) from sales) Example B modifies ytd_sales column to reflect the most recent sales recorded in the sales table. This assumes that only one set of sales is recorded for a given title on a given date and that updates are up-to-date. Options SET Specifies the value for a particular column name. The value can be an expression or a null value. When more than one column name and value pair is listed, they must be separated by commas. FROM Is required to pull data from other tables or views to modify rows in the table or view specified in the UPDATE clause. A standard FROM clause lists the tables and views referenced in the WHERE clause. WHERE A standard WHERE clause. Comments Use UPDATE to change values in rows that have already been inserted. Use INSERT to add full or partial new rows. UPDATE interacts with the IGNORE_DUP_KEY, IGNORE_DUP_ROW, and ALLOW_DUP_ROW statements set with the CREATE INDEX statement. (See "CREATE INDEX" for more information.) A rule can be created and bound to a column to restrict the domain of legal values that can be entered into it. Rules are created with CREATE RULE and bound with the sp_bindrule system procedure. You can define a trigger that takes a specified action when an UPDATE statement is executed on a specified table or on a specified column of a table. Permissions UPDATE permission defaults to the table or view owner, who can transfer it to other users. See Also CREATE INDEX, CREATE RULE, CREATE TRIGGER, INSERT, WHERE Clause, sp_bindefault, sp_bindrule, sp_unbindefault, sp_unbindrule UPDATE STATISTICS ──────────────────────────────────────────────────────────────────────────── Function Updates information about the distribution of key values in specified indexes. Syntax UPDATE STATISTICS table_name [[database.]owner.][index_name] Examples A. update statistics authors B. update statistics authors au_id_ind Comments SQL Server keeps statistics about the distribution of the key values in each index and uses these statistics in its decisions about which index(es) to use in query processing. Use the UPDATE STATISTICS statement when you create an index on already existing data or if a great deal of data in an indexed column has been added, changed, or removed (that is, if you suspect that the distribution of key values has changed). If you do not specify an index name, the statement updates the distribution statistics for all the indexes in the specified table. Since TRANSACT-SQL does not require index names to be unique in a database, you must give the name of the table with which the index is associated. Permissions UPDATE STATISTICS permission defaults to the table owner and is not transferable. See Also CREATE INDEX, sp_helpindex USE ──────────────────────────────────────────────────────────────────────────── Function Changes the current database. Syntax USE database_name Examples use pubs This example changes the current database to pubs. Comments USE must be in a batch by itself for you to reference objects in that database. An alias permits a user to use a database by an alias name. Use the sp_addalias system procedure. Permissions Permission for the USE statement is assigned when the Database Owner assigns permission by executing the sp_adduser system procedure. See Also CREATE DATABASE, DROP DATABASE, sp_addalias, sp_adduser, sp_defaultdb Variables (Local and Global) ──────────────────────────────────────────────────────────────────────────── Function Defined entities that are assigned values. A local variable is defined with a DECLARE statement and assigned an initial value within the statement batch where it is declared with a SELECT statement. Global variables are predefined. Syntax Local variable declaration: DECLARE @variable_name datatype [, @variable_name datatype...] Assignment of values to a local variable: SELECT @variable_name = expression [, @variable_name = expression...] [FROM clause] [WHERE clause] [GROUP BY clause] [HAVING clause] [ORDER BY clause] [COMPUTE clause] Examples A. declare @veryhigh money select @veryhigh = max(price) from titles if @veryhigh > $20 print "Ouch!" B. declare @one varchar(18), @two varchar(18) select @one = "this is one", @two = "this is two" if @one = "this is one" print "you got one" if @two = "this is two" print "you got two" else print "nope" Options @variable_name The name of a variable. Variable names must be preceded by the "at" symbol (@). The variable name conforms to the rules for identifiers except that it can contain only 29 characters. Variables can be used only in place of constants. They cannot be used in place of the names of tables, columns, other database objects, or keywords. datatype A system datatype or a user datatype. Comments Variables cannot be of the text or image datatype. Local variables are declared in the body of a procedure or query with the DECLARE keyword and given values with a SELECT statement. The "@" denotes a local variable name. Local variables are often used in a batch or procedure as counters for WHILE loops or IF...ELSE blocks. The SELECT statement that assigns a value to the local variable usually returns a single value. If the SELECT variable assignment statement returns more than one value, the variable is assigned the last value returned. The SELECT statement that assigns values to variables cannot be used to retrieve data. For example, the following is not legal: /* ILLEGAL STATEMENT ** declare @veryhigh money ** select @veryhigh = max(price), title_id ** from titles */ In addition to local variables, there are predefined global variables that can be used without being declared. Global variables are distinguished from local variables by having two @ signs preceding their names, for example, @ERROR, @ROWCOUNT. Many of the global variables report on system activity. Their reports refer to activity since the last time SQL Server was started. The global variables are as follows: @@CONNECTIONS The number of logins or attempted logins since SQL Server was last started. @@CPU_BUSY The amount of time, in ticks, that the CPU has spent doing SQL Server work since the last time SQL Server was started. @@ERROR The last error number generated by the system. @@IDLE The amount of time, in ticks, that SQL Server has been idle since it was last started. @@IO_BUSY The amount of time, in ticks, that SQL Server has spent doing input and output operations since it was last started. @@MAX_CONNECTIONS The maximum number of simultaneous connections that can be made with SQL Server in this computer environment. The user can configure SQL Server for fewer connections with sp_configure. @@NESTLEVEL The nesting level of the current execution (initially zero). Each time a stored procedure calls another stored procedure, the nesting level is incremented. If the maximum of 32 is exceeded, the transaction aborts. @@PACK_RECEIVED The number of input packets read by SQL Server since it was last started. @@PACKET_ERRORS The number of errors that have occurred while SQL Server was sending and receiving packets. @@PACK_SENT The number of output packets written by SQL Server since it was last started. @@PROCID The stored procedure ID of the currently executing procedure. @@ROWCOUNT The number of rows affected by the last statement (includes SELECT, INSERT, UPDATE, DELETE, the CREATE statements, and so on). @@TEXTSIZE The current value of the TEXTSIZE option of the SET statement, which specifies the length in bytes of text to be returned with a SELECT statement. @@TIMETICKS The number of microseconds per tick. The amount of time per tick is machine dependent. Each tick on the operating system is 21.25 milliseconds (1/32 second). @@TOTAL_ERRORS The number of errors that have occurred while SQL Server was reading or writing. @@TOTAL_READ The number of disk reads by SQL Server since it was last started (this includes disk reads only, not cache reads). @@TOTAL_WRITE The number of writes by SQL Server since it was last started. @@TRANCOUNT The number of currently active transactions for the current user. @@VERSION The date of the current version of SQL Server. The @ERROR global variable is commonly used to check the error status (succeeded or failed) of the most recently executed statement. A statement such as "if @ERROR != 0 return" causes an exit if an error had occurred. For information on the contents of many of these global variables, execute the sp_monitor system procedure. See Chapter 2, "System Procedures," for details. See Also DECLARE, Datatypes, PRINT, RAISERROR, sp_monitor Views ──────────────────────────────────────────────────────────────────────────── Function Provide an alternate way of looking at data in one or more tables. Syntax CREATE VIEW [owner.]view_name [(column_name [, column_name]...)] AS SELECT_statement Comments See "CREATE VIEW" for syntax, examples, explanations of keywords, and comments on creating a view. You cannot create a trigger on a view. If a view depends on a table (or view) that has been dropped, SQL Server produces an error message if anyone tries to use the view. If a new table (or view) is created to replace the one that has been dropped, the view becomes usable again. There are no restrictions on querying through views, but there are some restrictions on data update through views. Data update statements may not change any column in the view that is a computation. Data update statements referring to a view that includes aggregate functions (a built-in function and a GROUP BY or COMPUTE BY clause) are not allowed. INSERT statements are not allowed unless all NOT NULL columns in the underlying table or view are included in the view through which you are inserting new rows. (SQL Server has no way to supply values for NOT NULL columns in the underlying table or view.) INSERT and UPDATE statements are not allowed unless the columns being updated all belong to the same base table. (It's illegal in SQL to use data update statements on more than one table in a single statement.) However, if a view contains columns from more than one table but the data update statement references columns from only one of the tables, the statement is legal. When you query through a view, SQL Server checks to make sure that all the database objects referenced anywhere in the statement exist, that they are valid in the context of the statement, and that data update statements do not violate data integrity rules. If any of these checks fail, you get an error message. If the checks are successful, it "translates" the view into an action on the underlying table(s). Permissions You can use views as security mechanisms by granting permission on a view but not on one or more of its underlying tables. See Also CREATE VIEW, DROP VIEW, sp_depends, sp_help, sp_helptext, sp_rename WAITFOR ──────────────────────────────────────────────────────────────────────────── Function Specifies a specific time, a time interval, or an event for the execution of a statement block, stored procedure, or transaction. Syntax WAITFOR {DELAY "time" | TIME "time" | ERROREXIT | PROCESSEXIT} Examples A. begin waitfor time "14:20" insert chess(next_move) values('Q-KR5') execute sendmail 'judy' end At 2:20 p.m., the chess table will be updated with my next move, and a procedure called sendmail will insert a row in a table owned by Judy and notifies her that a new move now exists in the chess table. B. begin waitfor delay "00:00:10" print "Ten seconds have passed. Your time is up." end Options DELAY Instructs SQL Server to wait until the specified amount of time has passed, up to a maximum of 24 hours. TIME Instructs SQL Server to wait until the specified time. time A time in one of the acceptable formats for datetime data. You cannot specify dates─the date portion of the datetime value is not allowed. ERROREXIT Instructs SQL Server to wait until a kernel or user process terminates abnormally. PROCESSEXIT Instructs SQL Server to wait until a kernel or user process terminates for any reason. Comments After executing the WAITFOR statement, you cannot use your connection to SQL Server until the time or event that you specified occurs. You can use WAITFOR ERROREXIT with a procedure that kills the abnormally terminated process to free system resources that would otherwise be taken up by an infected process. To find out which process terminated, check the sysprocesses table with the sp_who system procedure. The time you specify with WAITFOR TIME or WAITFOR DELAY can include hours, minutes, and seconds. Use the format "hh:mi:ss", as described in "Datatypes." For example, WAITFOR TIME "16:23" instructs SQL Server to wait for 4:23 pm. The statement WAITFOR DELAY "01:30" instructs SQL Server to wait 1 hour and 30 minutes. For details on acceptable time formats, see "Datatypes." See Also BEGIN...END, Datatypes, sp_who WHERE Clause ──────────────────────────────────────────────────────────────────────────── Function Sets the conditions in a WHERE clause. (Joins and subqueries are specified in the search conditions; see the "Joins" and "Subqueries" sections for full details.) Syntax Search conditions immediately follow the keywords WHERE or HAVING in a SELECT, INSERT, UPDATE, or DELETE statement. (HAVING search conditions differ from WHERE search conditions only in that aggregate functions are not allowed in WHERE conditions.) If you use more than one of the search conditions in a single statement, connect the conditions with AND or OR. WHERE [NOT] expression comparison_operator expression WHERE [NOT] column_name [NOT] LIKE "match_string" WHERE [NOT] column_name IS [NOT] NULL WHERE [NOT] expression [NOT] BETWEEN expression AND expression WHERE [NOT] expression [NOT] IN ({value_list | subquery}) WHERE [NOT] EXISTS (subquery) WHERE [NOT] expression comparison_operator {ANY | ALL} (subquery) WHERE [NOT] column_name join_operator column_name WHERE [NOT] boolean_expression WHERE [NOT] expression {AND | OR} [NOT] expression Examples A. where advance * $2 > ytd_sales * price B. where phone not like '415%' Example B finds all the rows in which the phone number does not begin with 415. C. where au_lname like "[CK]ars[eo]n" Example C finds the rows for authors named Carson, Carsen, Karsen, and Karson. D. where advance < $5000 or advance is null E. where (type = "business" or type = "psychology") and advance > $5500 F. where ytd_sales between 4095 and 12000 G. where state in ('CA', 'IN', 'MD') Example G finds the rows in which the state is one of the three in the list. Options expression A column name; a constant; a function; any combination of column names, constants, and functions connected by arithmetic or bitwise operators; or a subquery. The arithmetic operators are as follows: Symbol Meaning ──────────────────────────────────────────────────────────────────────────── + Addition - Subtraction * Multiplication / Division % Modulo ──────────────────────────────────────────────────────────────────────────── Addition, subtraction, division, and multiplication can be used on int, smallint, tinyint, float, and money columns. The modulo operator cannot be used with money or float columns. A modulo is the integer remainder after a division involving two integers. For example, 21 % 9 = 3 because 21 divided by 9 equals 2 with a remainder of 3. The bitwise operators are as follows: Symbol Meaning ──────────────────────────────────────────────────────────────────────────── & Bitwise and (two operands) | Bitwise or (two operands) ^ Bitwise exclusive or (two operands) ~ Bitwise not (one operand) ──────────────────────────────────────────────────────────────────────────── The bitwise operators can be used on int, smallint, and tinyint columns only. All the bitwise operators translate the integer parameters into binary representation before evaluating them. (See "Expressions" for more information.) comparison_operator One of the following: Symbol Meaning ──────────────────────────────────────────────────────────────────────────── = Equal to > Greater than < Less than >= Greater than or equal to <= Less than or equal to != Not equal to !> Not greater than !< Not less than ──────────────────────────────────────────────────────────────────────────── In comparing char and varchar data, < means closer to the beginning of the alphabet and > means closer to the end of the alphabet. Case and special character evaluations depend on the collating sequence of the operating system on the machine on which SQL Server is located. On ASCII machines, for example, lowercase letters are greater than uppercase letters, and uppercase letters are greater than numbers. Trailing blanks are ignored for the purposes of comparison. So, for example, "Dirk" is the same as "Dirk ". In comparing dates, < means earlier and > means later. Put quotation marks around all character and datetime data used with a comparison operator. For example, = "Bennet", > "94609". (See "Datatypes.") column_name The name of a column used in the comparison. Qualify the column name with its table or view name if there is any ambiguity. NOT Negates any boolean expression and keywords, such as LIKE, NULL, BETWEEN, IN, and EXISTS. LIKE Indicates that the following character string (enclosed by single or double quotation marks) is a matching pattern. LIKE is available for char, varchar, and datetime columns (but not to search for seconds or milliseconds). You can use the LIKE keyword and wildcard characters with datetime data as well as with char and varchar. When you use LIKE with datetime values, SQL Server converts the dates to the standard datetime format and then to varchar. Since the standard storage format doesn't include seconds or milliseconds, you cannot search for seconds or milliseconds with LIKE and a pattern. It is a good idea to use LIKE when you search for datetime values, since datetime entries can contain a variety of date parts. For example, if you insert the value "9:20" into a column named arrival_time, the clause WHERE arrival_time = "9:20" would not find it because SQL Server converts the entry into "Jan 1, 1900 9:20AM." However, the clause WHERE arrival_time LIKE "%9:20%" would find it. match_string A string of characters and wildcards enclosed in quotation marks. For complete information, see "Wildcard Characters." The wildcards are as follows: Symbol Meaning ──────────────────────────────────────────────────────────────────────────── % Any string of 0 or more characters _ (underscore) Any single character [ ] Any single character within the specified range ([a-f]) or set ([abcdef]) [^] Any single character not within the specified range ([^a-f]) or set ([^abcdef]) ──────────────────────────────────────────────────────────────────────────── IS NULL Searches for null values. (=NULL is allowed in UPDATE statements. Outside of this one exception, NULL returns only NULL when used with arithmetic, bitwise, or comparison operators. NULL is essentially useless as an operand with any of these operators.) BETWEEN The range-start keyword. Use AND for the range-end value. A range of BETWEEN x AND y, unlike a range of > x < y, is inclusive. IN Allows you to select values that match any one of a list of values. The comparator can be a constant or a column name, and the list can be a set of constants or, more commonly, a subquery. (See "Subqueries" for information on using IN with a subquery.) Enclose the list of values in parentheses. value_list A list of values. Put single or double quotation marks around character values, and separate each value from the following one with a comma. (See example E, earlier in this section.) EXISTS Is used with a subquery to test for the existence of some result from the subquery. (See "Subqueries" for more information.) subquery A restricted SELECT statement (ORDER BY and COMPUTE clauses and the INTO keyword are not allowed) inside the WHERE or HAVING clause of a SELECT, INSERT, DELETE, UPDATE, or subquery. (See "Subqueries" for more information.) ANY Is used with >, <, or = and a subquery. It returns results when any value retrieved in the subquery matches the value in the WHERE or HAVING clause of the outer statement. (See "Subqueries" for more information.) ALL Is used with > or < and a subquery. It returns results when all values retrieved in the subquery match the value in the WHERE or HAVING clause of the outer statement. (See "Subqueries" for more information.) join_operator A comparison operator, =*, or *= . (See "Joins" for more information.) boolean_expression An expression that returns true or false. (See "Expressions" for a full definition.) AND Joins two conditions and returns results when both of the conditions are true. When more than one logical operator is used in a statement, AND operators are normally evaluated first. However, you can change the order of execution with parentheses. OR Joins two conditions and returns results when either of the conditions is true. When more than one logical operator is used in a statement, OR operators are normally evaluated after AND operators. However, you can change the order of execution with parentheses. Comments WHERE and HAVING search conditions are identical, except that aggregate functions are not permitted in WHERE clauses ("HAVING avg(price) > 20" is legal; "WHERE avg(price) > 20" is not). See "Aggregate Functions" for information on the use of aggregate functions. If there are columns in the select list that neither have aggregate functions applied to them nor are included in the query's GROUP BY clause (illegal in standard SQL), the meanings of HAVING and WHERE are somewhat different. In this situation, a WHERE clause restricts the rows that are included in the calculation of the aggregate function but does not restrict the rows returned by the query. Conversely, a HAVING clause restricts the rows returned by the query but does not affect the calculation of the aggregate function. See "GROUP BY and HAVING Clauses" for examples. There are two ways to specify literal quotation marks within a char or varchar entry. The first method is to use two quotation marks. For example, if you have begun a character entry with a single quotation mark and wish to include a single quotation mark as part of the entry, use two single quotation marks: 'I don"t understand.' With double quotation marks: "He said, ""It's not really confusing.""" The second method is to enclose a quote in the opposite kind of quotation mark. In other words, surround an entry containing a double quotation mark with single quotation marks (or vice versa). Here are some examples: 'George said, "There must be a better way."' "Isn't there a better way?" 'George asked, "Isn"t there a better way?"' To enter a character string longer than the width of your screen, enter a backslash (\) before going to the next line. See Also Datatypes, DELETE, EXECUTE, Expressions, Functions, INSERT, Joins, Search Conditions, SELECT, Subqueries, UPDATE, Wildcard Characters, sp_helpjoins WHILE ──────────────────────────────────────────────────────────────────────────── Function Sets a condition for the repeated execution of a statement or statement block. The statement(s) is executed repeatedly as long as the specified condition is true. Syntax WHILE boolean_expression statement Examples while (select avg(price) from titles) < $30 begin update titles set price = price * $2 select max(price) from titles if (select max(price) from titles) > $50 break else continue end begin select title_id, price from titles print "Too much for the market to bear" end If the average price is less than $30, double the prices. Then select the maximum price. If it is less than or equal to $50, restart the WHILE loop and double the prices again. If the maximum price is more than $50, exit the WHILE loop, select the title_ids and the too-high prices, and print a message. Comments The execution of statements in the WHILE loop can be controlled from inside the loop with the BREAK and CONTINUE statements. CONTINUE causes the WHILE loop to restart, skipping any statements after CONTINUE. BREAK causes an exit from the WHILE loop, and any statements that appear after the END keyword that marks the end of the loop are executed. BREAK and CONTINUE are often (but not always) activated by IF tests. If two or more WHILE loops are nested, BREAK exits to the next outermost loop. First all the statements after the end of the inner loop run, and then the next outermost loop restarts. See Also BEGIN...END, BREAK, CONTINUE, CREATE PROCEDURE, GOTO Wildcard Characters ──────────────────────────────────────────────────────────────────────────── Function Used with the LIKE keyword to represent any character in a string when searching for a char, varchar, or datetime value. Syntax {WHERE | HAVING} [NOT] expression [NOT] LIKE "string" The string can include these wildcard characters: Wildcard Meaning ──────────────────────────────────────────────────────────────────────────── % Any string of 0 or more characters _ (underscore) Any single character [ ] Any single character within the specified range ([a-f]) or set ([abcdef]) [^] Any single character not within the specified range ([^a-f]) or set ([^abcdef]) ──────────────────────────────────────────────────────────────────────────── Examples A. select phone from authors where phone like "415%" Example A finds all the phone numbers in the authors table that begin with the 415 area code. B. select phone from authors where phone not like "415%" select phone from authors where not phone like "415%" These two queries are equivalent: they find all the phone numbers in the authors table that do not begin with the 415 area code. C. select au_lname from authors where au_lname like "%en%" Example C finds names that have the characters "en" in them (Bennet, Green, McBadden). D. select au_fname from authors where au_fname like "_heryl" Example D finds six-letter names that end with "heryl" (Cheryl). E. select au_lname from authors where au_lname like "[M-Z]inger" Example E finds names ending with "inger" and beginning with any single character between "M" and "Z". F. select au_lname from authors where au_lname like "[dD]eFrance" Example F finds both DeFrance and deFrance. G. select au_lname from authors where au_lname like "M[^c]%" Example G finds names beginning with "M" whose second letter is not "c". Comments The string with wildcard characters must always be enclosed by single or double quotation marks. LIKE and the wildcard characters can be used for searching char, varchar, and datetime columns (but not to search for seconds or milliseconds). Wildcards used without LIKE have no special meaning. For example, this query finds any phone numbers that start with the four characters "415%": select phone from authors where phone = "415%" You can use the LIKE keyword and wildcard characters with datetime data as well as with char and varchar. When you use LIKE with datetime values, SQL Server converts the dates to the standard datetime format and then to varchar. Since the standard storage format doesn't include seconds or milliseconds, you cannot search for seconds or milliseconds with LIKE and a pattern. It is a good idea to use LIKE when you search for datetime values, since datetime entries can contain a variety of date parts. For example, if you insert the value "9:20" into a column named arrival_time, the clause WHERE arrival_time = "9:20" would not find it because SQL Server converts the entry into "Jan 1, 1900 9:20AM". However, the clause WHERE arrival_time LIKE "%9:20%" would find it. To use %, _, [ ], or [^ ] as literal characters in a LIKE string rather than as wildcards, use square brackets around the percent sign, the underscore, and the open bracket. Use the close bracket by itself. Use the dash as the first character inside a set of square brackets. ╓┌─────────────────┌─────────────────────────────────────────────────────────╖ Symbol Meaning ──────────────────────────────────────────────────────────────────────────── LIKE "5%" 5 followed by any string of 0 or more characters LIKE "5[%]" 5% LIKE "_n" an, in, on, and so on LIKE "[_]n" _n LIKE "[a-cdf]" a, b, c, d, or f LIKE "[-acdf]" -, a, c, d, or f LIKE "[[]" [ Symbol Meaning ──────────────────────────────────────────────────────────────────────────── LIKE "[[]" [ LIKE "]" ] ──────────────────────────────────────────────────────────────────────────── You can't always duplicate NOT LIKE patterns with LIKE and the negative wildcard [^] because NOT LIKE finds the items that do not match the entire LIKE pattern, while LIKE with negative wildcards is evaluated one character at a time. For example, to see all the system tables in a database, use this query since all system tables begin with these three letters: select name from sysobjects where name LIKE "sys%" To see all the objects that are not system tables, use NOT LIKE "sys%". If you have a total of 32 objects and the LIKE finds 13 names that match the pattern, the NOT LIKE will find exactly the 19 objects that do not match the LIKE pattern. You won't always find the same names with a pattern such as LIKE "[^s][^y][^s]%". Instead of 19, you might get only 14, with all the names that begin with "s" or have "y" as the second letter or have "s" as the third letter eliminated from the results as well as the system table names. This is because matchstrings with negative wildcards are evaluated in steps, one wildcard at a time. If the match fails at any point in the evaluation, it is eliminated. See Also CREATE PROCEDURE, EXECUTE, Expressions, Search Conditions, WHERE Clause WRITETEXT ──────────────────────────────────────────────────────────────────────────── Function Permits nonlogged, interactive updating of an existing text field. Syntax WRITETEXT [[database.]owner.]table_name.column_name text_ptr [WITH LOG] data Examples declare @val varbinary(30) select @val = textptr(blurb) from texttest where title_id = "BU7832" writetext texttest.blurb @val "hello world" Options table_name.column_name The name of the text column and the table name, which must be included. The database name and owner name are optional. WITH LOG Logs the inserted text or image data. This option aids media recovery, but text data quickly increases the size of the transaction log. Make sure that the transaction log resides in a separate database device. (See "sp_logdevice" in Chapter 2 and the SQL Server Administrator's Guide for details.) Comments By default, WRITETEXT is a nonlogged operation. This means that text or image data is not logged when it is written into the database. To use WRITETEXT in its default, nonlogged state, the System Administrator must use sp_dboption to set select into/bulkcopy. This permits the insertion of nonlogged data. WRITETEXT updates text data in an existing row. The update completely replaces existing text. For a valid text pointer to exist, a text column must contain either text data or a null value that has been explicitly entered with INSERT or UPDATE. Given a table textnull with columns key and x, where x is a text column that permits nulls: This insert assigns a valid text pointer: insert textnull values (1, null) go This insert does not assign a pointer: insert textnull (key) values (2) go You can update a null text column to get a valid text pointer or insert entirely new rows with INSERT (which is a logged operation) or bcp. The maximum length of text that can be inserted interactively with WRITETEXT is approximately 120K bytes for text data and 60K bytes for image data. You cannot use WRITETEXT on text and image columns in views. See Also READTEXT, Text/Image Datatypes, Text/Image Functions Chapter 2 System Procedures ──────────────────────────────────────────────────────────────────────────── This chapter consists of reference pages, in alphabetical order, for system procedures (SQL-Server-supplied stored procedures). Note that system procedure names always begin with "sp_". Each system procedure and its function is listed below: ──────────────────────────────────────────────────────────────────────────── sp_addalias Maps one user to another in a database. sp_addgroup Adds a group to a database. sp_addlogin Authorizes a new SQL Server user by adding an entry to master.dbo.syslogins. sp_addtype Creates a user-defined datatype. sp_addumpdevice Adds a dump device to SQL Server. sp_adduser Adds a new user to the current database. sp_bindefault Binds a default to a column or to a user-defined datatype. sp_bindrule Binds a rule to a column or user datatype. sp_changedbowner Changes the owner of a database. sp_changegroup Changes a user's group. sp_commonkey Defines a common key between two tables or views. sp_configure Displays or changes configuration options. sp_dboption Displays or changes database options. sp_defaultdb Changes a user's default database. sp_depends Displays information about database object dependencies. sp_diskdefault Sets a database device's status to defaulton or defaultoff. sp_dropalias Removes the alias username identity that had been established with sp_addalias. sp_dropdevice Drops an SQL Server database device or dump device. sp_dropgroup Drops a group from a database. sp_dropkey Removes a key from the syskeys table that had been defined using sp_primarykey, sp_foreignkey, or sp_commonkey. sp_droplogin Drops an SQL Server user by deleting his or her entry in master.dbo.syslogins. sp_droptype Drops a user-defined datatype by deleting the type from systypes. sp_dropuser Drops a user from the current database by deleting the entry from sysusers. sp_foreignkey Defines a foreign key on a table or view. sp_help Reports information about a database object (any object listed in sysobjects) or about an SQL Server-supplied or user-defined datatype. sp_helpdb Reports information about a database or about all databases. sp_helpdevice Reports information about SQL Server's database devices and dump devices. sp_helpgroup Reports information on a group or on all groups in the current database. sp_helpindex Reports index information on a table. sp_helpjoins Lists the columns in two tables or views that are likely to be joined. sp_helpkey Reports information on primary, foreign, and common keys. sp_helprotect Reports permissions by database object and optionally by user for that object. sp_helpsql Provides syntax for SQL statements and system procedures. sp_helptext Prints the text of a stored procedure, trigger, view, default, or rule. sp_helpuser Reports information on users of a database. sp_lock Reports information on locks. sp_logdevice Puts the syslogs system table, which contains the transaction log, on a separate database device. sp_monitor Displays statistics about SQL Server. sp_password Adds or changes a password for an SQL Server login ID. sp_primarykey Defines a primary key on a table or view. sp_rename Changes the name of a user-created object in the current database. sp_renamedb Changes the name of a database. sp_spaceused Displays the number of rows, the number of data pages, and the disk space used by an object or by each object in a database. sp_unbindefault Unbinds a default value from a column or from a user-defined datatype. sp_unbindrule Unbinds a rule from a column or user-defined datatype. sp_who Reports information on current SQL Server users and processes. ──────────────────────────────────────────────────────────────────────────── There are additional system procedures that are called by other system procedures; users cannot call them directly. sp_addalias ──────────────────────────────────────────────────────────────────────────── Function Maps one user to another in a database. Syntax sp_addalias login_id, username Examples sp_addalias victoria, albert There is a user named albert in the database's sysusers table and a login ID for a user named victoria in master.dbo.syslogins. This system procedure allows victoria to use the current database and to be known as albert in this database. Options login_id The master.dbo.syslogins name of the user who wants an alternate identity in the current database. An alias allows an SQL Server user to be known in a database as another user. If the user with login_id exists in the database's sysusers table, SQL Server won't find the user's alias identity, since it checks sysusers before checking sysalternates. username The name of the database user to whom the first user wishes to be linked. The name must exist in both master.dbo.syslogins and in the sysusers table of the current database Comments Executing sp_addalias maps one user (login_id) to another (username) in the current database. The mapping is shown in sysalternates, where the two users' suids are connected. A report on any users mapped to a specified user can be generated with sp_helpuser if you give the specified user's name as a parameter. When a user tries to use a database, SQL Server checks sysusers to see if the username is listed there. If the username is not there, it then checks sysalternates. If an entry (the user's suid) is found there, mapped to a database user's suid (altsuid in sysalternates table), the first user is treated as the second user while using the database. There are three ways that you can be authorized to use a database (assuming you are listed in master.dbo.syslogins). First, you can be listed in the sysusers table of that database. (New database users are added by the Database Owner with the sp_adduser system procedure.) Second, you can have an alias in the database, which has been added by the Database Owner with the sp_addalias system procedure. Third, you can use the database as guest if there is a guest entry in the database's sysusers table. Messages Alias user added. The procedure was successful. Now the user with login_id can use the current database. When doing so, he or she will be known as username. No login exists with the specified login name. There is no entry in master.dbo.syslogins for login_id. Everyone using SQL Server, whether aliased or not, must have a login ID. No user with the specified name exists in the current database. Since username is not a user in the database, login_id can't be aliased to it. The specified user name is already aliased. The login_id is already aliased to a user in the current database. A login_id may be aliased to only one database user at a time. To change an alias, first drop the current alias using sp_dropalias, then add the new alias. Permissions Execute permission defaults to the Database Owner. Tables Used master.dbo.syslogins, sysalternates, sysusers See Also sp_addlogin, sp_adduser, sp_dropalias, sp_helpuser, USE sp_addgroup ──────────────────────────────────────────────────────────────────────────── Function Adds a group to a database. Groups are used as collective names in granting and revoking permissions. Syntax sp_addgroup grpname Examples sp_addgroup accounting This example creates an accounting group. Options grpname The name of the group. Groupnames must follow the rules for identifiers. Comments Groups are used as collective names for granting and revoking permissions. The sp_addgroup system procedure adds an entry for the new group to a database's sysusers table. Each group's uid is a number greater than or equal to 16,384 (except the uid of the public group, which is always 0). Once a group has been created, you can use the groupname as an optional parameter to sp_adduser to add the new user to a group at the same time. To add an already existing user to a group, use sp_changegroup. A user can be a member of only one group. Every database is created with a public group. If a user is not explicitly added to another group, he or she is a member of the public group. Messages New group added. An entry for the group has been added to the current database's sysusers table. Users may now be made members of the group with the sp_adduser or sp_changegroup system procedures. A user with the specified group name already exists. The groupname you supplied is already being used as a username. Choose another name to use as the grpname parameter. A group with the specified name already exists. The groupname you supplied is already being used as a groupname. Choose another name to use as the grpname parameter. Permissions Execute permission defaults to the Database Owner. Tables Used sysusers See Also sp_adduser, sp_changegroup, sp_dropgroup, sp_helpgroup, GRANT, REVOKE sp_addlogin ──────────────────────────────────────────────────────────────────────────── Function Authorizes a new SQL Server user by adding an entry to master.dbo.syslogins. Syntax sp_addlogin login_id [, passwd [, defdb]] Examples A. sp_addlogin victoria Example A creates an SQL Server login for the user victoria. Since the passwd and defdb parameters are not included, victoria's password is the null value and her default database is master. B. sp_addlogin albert, food, corporate Example B creates an SQL Server login ID for the user albert. His password is food and his default database is corporate. Options login_id The login ID of the user. Login IDs normally follow the rules for identifiers. (However, numeric login IDs of up to 30 digits are also allowed.) passwd The user's password. If no password is given, the default password is the null value. defdb The name of the user's default, or "home," database─the database to which the user is connected when he or she logs in. If no defdb is given, the default is master. Comments After sp_addlogin is executed, the user can log in to SQL Server. The user can change his or her own password at any time with sp_password. The System Administrator can change any user's password with sp_password. If a third parameter (the name of a default database) is supplied, the user is connected to the specified database without the USE statement and without passing through master. However, the user cannot use the default database until he or she is given access to that database by its Database Owner (using the sp_adduser or sp_addalias system procedures). The user can change his or her own default database at any time with sp_defaultdb. The System Administrator can change any user's default database with sp_defaultdb. Messages New login created. An SQL Server login ID for login_id has been created successfully. Only the System Administrator may execute this procedure. Someone other than the System Administrator tried to execute the system procedure. Only the System Administrator can add login IDs to SQL Server. A user with the specified login name already exists. A user with that login ID has already been added. Choose another login_id. Database name not valid -- login not added. The specified default database does not exist. Create the database first or choose a database that already exists as defdb. Permissions Execute permission defaults to the System Administrator. Tables Used master.dbo.sysdatabases, master.dbo.syslogins See Also sp_addalias, sp_adduser, sp_defaultdb, sp_droplogin, sp_password, USE sp_addtype ──────────────────────────────────────────────────────────────────────────── Function Creates a user-defined datatype. Syntax sp_addtype typename, phystype[(length)] [, nulltype] Examples A. sp_addtype ssn, "varchar(11)" Example A creates a user-defined datatype called ssn to be used for columns that hold social security numbers. Since no nulltype parameter is specified, no null values are allowed. Notice that "varchar(11)" is enclosed in quotation marks because it contains punctuation (parentheses). B. sp_addtype birthday, datetime, null Example B creates a user-defined datatype called birthday that allows null values. Options typename The name of the user datatype. Type names must follow the rules for identifiers and must be unique for each owner in each database. phystype[(length)] The physical or SQL-Server-supplied type (char, int, and so on) on which the user datatype is based. Length must be specified for char, varchar, binary, and varbinary. If a length is given, the phystype parameter must be enclosed in single or double quotation marks. (Quotation marks are required around all parameters with embedded blank spaces or punctuation marks.) nulltype A variable that indicates how the user datatype handles null value entries. The default (if this optional parameter is omitted) is NOT NULL (disallow null values and require an explicit entry). Acceptable values for this parameter are NULL, NONULL, and NOT NULL. You can override the user datatype's nulltype condition in a CREATE TABLE statement. Comments Executing sp_addtype creates a user-defined datatype and adds it to the systypes system table. Once a user datatype is created, you can use it in CREATE TABLE statements and bind defaults and rules to it. Define each user datatype in terms of one of the physical (SQL-Server-supplied) datatypes, optionally specifying NULL (allow null entries) or NONULL (disallow them). The default, if you give no third parameter, is nonull. A user-defined datatype name must be unique in the database, but user-defined datatypes with different names can have the same definitions. Messages Type added. The new user datatype was created and can now be used in CREATE TABLE statements or to bind rules and defaults. Usage: sp_addtype name, 'datatype' [,null | nonull] Syntax summary. The nulltype parameter can be null, NULL, nonull, NONULL, not null, or NOT NULL. A type with the specified name already exists. The typename parameter is already a user-defined datatype. Choose a different name. Illegal length specified -- must be between 1 and 255. The length of a datatype must be between 1 and 255. Physical datatype does not exist. The phystype you gave is not an SQL Server datatype. Physical datatype does not allow null values. The bit datatype doesn't allow null values, and you specified with the nulltype parameter that you wanted to allow null values. Physical type is fixed length. You cannot specify the length. The physical datatypes that take length specifications are char, varchar, binary, and varbinary. Other physical datatypes have fixed lengths that cannot be changed. You must specify a length with this physical type. You used a phystype ─char, varchar, binary, or varbinary─that requires a length. For example, "char(10)" is acceptable, but char is not. Permissions Execute permission defaults to the public group. Tables Used systypes See Also sp_bindefault, sp_bindrule, sp_droptype, sp_rename, sp_unbindefault, sp_unbindrule, CREATE DEFAULT, CREATE RULE, CREATE TABLE sp_addumpdevice ──────────────────────────────────────────────────────────────────────────── Function Adds a dump device to SQL Server. Syntax sp_addumpdevice {"disk | diskette"}, logical_name, physical_name, cntrltype [, noskip | skip [, media_capacity]] Examples sp_addumpdevice "diskette", "diskettedumpf", "f:\sqlserver\dump\dsk.dat", 3, noskip, .36 Options disk", "diskette" The type of dump device. Use "disk" for adding a hard disk file as a dump device. Use "diskette" for adding a 5-1/4 inch or 3-1/2 inch diskette drive file as a dump device. These options must be enclosed in quotation marks. logical_name The logical name of the dump device, which is used in the LOAD and DUMP SQL statements. This name must follow the rules for identifiers. When you add a dump device with a logical name longer than 30 characters, the dump device is created but the name is truncated to 30 characters. physical_name The physical name of the dump device. This name must follow the rules for operating system filenames and must include a full pathname. cntrltype The controller number of the new dump device. The same controller number can be used for more than one dump device. Different numbers are necessary only if you want to use more than one device simultaneously. For hard disk dump devices, the controller number parameter must be 2. For diskette dump devices, the controller number must be 3 or 4. If there are two diskette dump devices, use controller number 3 for one and controller number 4 for the other. skip, noskip Indicates whether ANSI tape labels are read (noskip) or ignored (skip). Use noskip for diskette dump devices. (The skip option is reserved for future tape support.) media_capacity Diskette size in megabytes. For diskette devices, 1.2 megabytes is the default. If you specify media_capacity, you must also specify the noskip option. The console program prompts the operator to change diskettes when a database dump reaches the specified capacity of the diskette. Comments The sp_addumpdevice system procedure adds a dump device to the master.dbo.sysdevices table. It can then be referred to logically in the DUMP DATABASE, DUMP TRANSACTION, LOAD DATABASE, and LOAD TRANSACTION statements. To add database devices to sysdevices, use the DISK INIT statement. ──────────────────────────────────────────────────────────────────────────── Warning If you add a dump device with a physical name of NULL and any logical name, the dump device will be added. However, its physical name will be changed to nul and its logical name changed to diskdump. ──────────────────────────────────────────────────────────────────────────── Messages Unknown device type. Use 'disk' or 'diskette'. The value supplied for the first parameter isn't a known dump device type. Device with same logical name already exists. All SQL Server dump devices must have unique logical names. There is already a dump device with the name supplied for the logical_name parameter. Diskette device added. The diskette dump device was added successfully. Disk device added. The disk dump device was added successfully. For disk devices, controller number parameter must be 2. When you are adding a disk dump device, the value for the fourth parameter, cntrltype, must be 2. For diskette devices, controller number parameter must be 3 or 4. When you are adding a diskette dump device, the value for the fourth parameter, cntrltype, must be 3 or 4. Please specify media capacity in megabytes (1 megabytes minimum). For diskette dump devices, enter the size in megabytes. The default is 1.2 megabytes. The console program prompts the operator to change diskettes when a database dump reaches the specified capacity of the diskette. Permissions Execute permission defaults to the System Administrator. Tables Used master.dbo.sysdevices See Also sp_helpdevice, sp_dropdevice, DISK INIT, DUMP DATABASE, DUMP TRANSACTION, LOAD DATABASE, LOAD TRANSACTION, console sp_adduser ──────────────────────────────────────────────────────────────────────────── Function Adds a new user to the current database. Syntax sp_adduser login_id [, username [, grpname]] Examples A. sp_adduser victoria, victoria, fort_mudge SQL Server login ID victoria has already been added to SQL Server with sp_addlogin. In example A we add her to the database. Victoria will be known in the current database by the username victoria and belong to the group fort_mudge (which has previously been created with sp_addgroup). B. sp_adduser margaret Example B adds margaret to the database. Her database username is the same as her SQL Server login ID, and she belongs to the default group public. C. sp_adduser haroldq, harold, fort_mudge Example C adds haroldq to the database. When haroldq uses the current database, his name is harold. He belongs to the fort_mudge group. Options login_id The user's login ID as found in master.dbo.syslogins. username A name for this user in the current database. Omitting this optional parameter causes the username to default to the user's login ID. grpname The name of an existing group in the database. If you don't specify a group, the user becomes a member of the default group public. If you want to add a user to a group, you must supply a value for the username parameter. It can be the same as login_id. Comments The Database Owner executes sp_adduser to add a username to the sysusers table of the current database, enabling the user to access the current database under his or her own name. Specifying a username parameter gives the new user a name in the database different from his or her login ID on SQL Server. The ability to assign a user a different name on SQL Server and in a database is provided as a convenience. This is not an alias as provided by sp_addalias: the user is simply given a different name and is not mapped to the identity and permissions of another user. You can also supply a grpname for the user. The group must already exist in the database. A user can be a member of only one group. Every user is a member of the default group, public, if he or she has not been explicitly added to some other group with sp_adduser or sp_changegroup. To access a database, a user must be listed in sysusers (with sp_adduser) or mapped to another user in sysalternates (with sp_addalias), or there must be a guest entry in sysusers. Messages New user added. The system procedure was successful. The user is now known in the current database. No login with the specified name exists. The login_id you gave is unknown to SQL Server. Each user must have a login ID on SQL Server before he or she can be added to a database. A user with the same name already exists in the database. The username is already a user in the database. Choose another name. User already has alias access to the database. The login_id is already known to the database by an alias. If you still want to add the user, drop the alias with the sp_dropalias system procedure and then re-execute sp_adduser. User already has a login under a different name. The user with the login_id you supplied is listed in the current database's sysusers table with a name different from the one supplied as the username parameter. No group with the specified name exists. The groupname you supplied does not exist in this database. Either omit the grpname parameter or create the group with sp_addgroup first. Permissions Execute permission defaults to the Database Owner. Tables Used master.dbo.syslogins, sysalternates, sysusers, See Also sp_addalias, sp_addgroup, sp_changegroup, sp_dropalias, sp_dropgroup, sp_helpuser, GRANT, REVOKE, USE sp_bindefault ──────────────────────────────────────────────────────────────────────────── Function Binds a default to a column or to a user-defined datatype. Syntax sp_bindefault defname, objname [, futureonly] Examples A. sp_bindefault today, "employees.startdate" Assuming that a default named today has been defined in the current database with CREATE DEFAULT, the system procedure in example A binds it to the startdate column of the employees table. Whenever a row is added to the employees table and data for the startdate column is not supplied, the column gets the value of the default today. B. sp_bindefault def_ssn, ssn Assuming that a default named def_ssn and a user-defined datatype named ssn exist, the system procedure in example B binds def_ssn to ssn. The default is inherited by all columns that are assigned the user datatype ssn when a table is created. Existing columns of type ssn also inherit the default def_ssn unless you use the optional third parameter, futureonly (which prevents existing columns of that user datatype from inheriting the default), or unless the column's default has previously been changed (in which case the changed default is maintained). C. sp_bindefault def_ssn, ssn, futureonly Example C binds the default def_ssn to the user datatype ssn. Because the futureonly parameter is included, no existing columns of type ssn are affected. Options defname The name of a default. Defaults are created with CREATE DEFAULT statements and then bound to specific columns or user-defined datatypes with sp_bindefault. objname The name of the table and column or user-defined datatype to which the default is to be bound. If the objname is not of the form "table.column", it is assumed to be a user datatype. (Quotation marks are required around all system procedure parameters that have embedded blanks or punctuation.) By default, existing columns of the user datatype inherit the default defname unless the column's default has previously been changed. futureonly Prevents existing columns of a user datatype from inheriting the new default. This parameter is optional when binding a default to a user-defined datatype. It is never used when binding a default to a column. Comments First use the CREATE DEFAULT statement to create a default. Then execute sp_bindefault to bind it to a column or user datatype in the current database. You cannot bind a default to an SQL-Server-supplied datatype. If binding to a column, the objname parameter must be of the form "table.column". Any other format is taken to be the name of a user datatype. You can bind a new default to a column or user datatype using sp_bindefault without unbinding an existing default. The old default is overridden. Existing columns of the user-defined datatype inherit the new default unless their default has previously been changed or the value of the optional third parameter is futureonly. New columns of the user-defined datatype always inherit the default. The default must be compatible with the datatype of the column. You cannot use "N/A", for example, as a default for a numeric column. If the default is not compatible with the column to which you've bound it, SQL Server will generate an error message when it tries to insert the default value (not when you bind it). Messages Default bound to column. The default was successfully bound to the specified column in the specified table. Default bound to datatype. The default was successfully bound to the specified user-defined datatype. Default and table or usertype must be in 'current' database. The objname parameter supplied with the system procedure contained a reference to another database. Defaults can be bound to objects in the current database only. No such default exists. You must create the default first. First create the default in the current database with CREATE DEFAULT. Then execute sp_bindefault. You do not own a column of that name. Only the owner of a table can bind a default to any of its columns. You are not the owner. You do not own a datatype of that name. Only the owner of a user-defined datatype can bind a default to it. You are not the owner. The new default has been bound to column(s) of the specified user datatype. Existing columns of the user-defined datatype specified now have the new default bound to them (unless their defaults were previously changed). Permissions Execute permission defaults to the object owner. Tables Used syscolumns, sysobjects, systypes See Also sp_bindrule, sp_unbindefault, sp_unbindrule, CREATE DEFAULT, DROP DEFAULT sp_bindrule ──────────────────────────────────────────────────────────────────────────── Function Binds a rule to a column or user datatype. Syntax sp_bindrule rulename, objname [, futureonly] Examples A. sp_bindrule today, "employees.startdate" Assuming that a rule named today has been created in the current database with CREATE RULE, the system procedure in example A binds it to the startdate column of the employees table. When a row is added to employees, the data for the startdate column is checked against the today rule. B. sp_bindrule rule_ssn, ssn Assuming the existence of a rule named rule_ssn and a user datatype named ssn, the system procedure in example B binds rule_ssn to ssn. In a CREATE TABLE statement, columns of type ssn inherit the rule_ssn rule. Existing columns of type ssn also inherit the rule_ssn rule unless you use the optional third parameter, futureonly (which prevents existing columns of type ssn from inheriting the rule) or ssn's rule has previously been changed (in which case the changed rule is maintained). C. sp_bindrule rule_ssn, ssn, futureonly The rule_ssn rule is bound to the user type ssn, but no existing columns of type ssn are affected (example C). Options rulename The name of a rule. Rules are created with CREATE RULE statements and then bound to specific columns or user datatypes with sp_bindrule. objname The name of the table and column or user datatype to which the rule is to be bound. If the parameter is not of the form "table.column", it is assumed to be a user datatype. (Quotation marks are required around all system procedure parameters that have embedded blanks or punctuation.) futureonly Prevents existing columns of a user datatype from inheriting the new rule. This parameter is optional when binding a rule to a user datatype. It is never used when binding a rule to a column. Comments The rule is enforced when an INSERT is attempted, not at binding. You can bind a character rule to a column of numeric datatype, even though such an INSERT is illegal. First use the CREATE RULE statement to create a rule. Then execute sp_bindrule to bind it to a column or user datatype in the current database. You cannot bind a rule to an SQL-Server-supplied datatype. If binding to a column, the objname parameter must be of the form "table.column". Any other format is taken to be the name of a user datatype. You can bind a new rule to a column or user datatype using sp_bindrule without unbinding an existing rule. The old rule is overridden. Existing columns of the user datatype inherit the new rule unless their rule has previously been changed or the value of the optional third parameter is futureonly. New columns of the user datatype always inherit the rule. Messages Rule bound to table column. The rule was successfully bound to the specified column in the specified table. Rule bound to datatype. The rule was successfully bound to the specified user datatype. Rule and table or usertype must be in 'current' database. The objname parameter supplied with the procedure contained a reference to another database. Rules can be bound to objects in the current database only. No such rule exists. You must create the rule first. First create the rule in the current database with CREATE RULE. Then execute sp_bindrule. You do not own a column of that name. Only the owner of a table can bind a rule to any of its columns. You are not the owner. You do not own a datatype of that name. Only the owner of a user datatype can bind a rule to it. You are not the owner. The new rule has been bound to column(s) of the specified user datatype. Existing columns of the specified user datatype now have the new rule bound to them (unless their rules were previously changed). Permissions Execute permission defaults to the object owner. Tables Used syscolumns, sysobjects, systypes See Also sp_bindefault, sp_unbindefault, sp_unbindrule, CREATE RULE, DROP RULE sp_changedbowner ──────────────────────────────────────────────────────────────────────────── Function Changes the owner of a database. Syntax sp_changedbowner login_id Examples sp_changedbowner albert This example makes the user albert the owner of the current database. Options login_id The login ID of the new owner of the current database. The new owner must not already be known as either a user or alias (that is, the new owner must not already be listed in sysusers or sysalternates). Comments The sp_changedbowner system procedure makes login_id the owner of the current database. The new owner must already have a login ID on SQL Server but must not have a database name or alias name in the database. If you want to assign database ownership to such a user, drop his or her database name and alias entries before executing sp_changedbowner. After the sp_changedbowner system procedure is executed, the new owner is known as Database Owner inside the database. If you (as System Administrator) want to grant permissions to this new owner, grant them to Database Owner because the user is no longer known inside the database under any other name. Messages Only the System Administrator can change the owner of a database. You are not the System Administrator, so you can't change the owner of the current database. You can't change the owner of the Master Database. The owner of the master database cannot be changed by anyone. No login with the specified name exists. The proposed new Database Owner must have a login ID on SQL Server. The proposed new db owner already is a user in the database. The specified login_id is already a user in the current database. To make him or her the Database Owner, first drop his or her user entry from the current database's sysusers table. The proposed new db owner already is aliased in the database. The specified login_id is already aliased in the current database. To make him or her the Database Owner, first drop his or her user alias entry from the current database's sysalternates table. Database owner changed. The Database Owner was successfully changed. Permissions Execute permission defaults to the Database Owner. Tables Used master.dbo.sysdatabases, master.dbo.syslogins, sysalternates, sysusers See Also sp_addlogin, sp_dropalias, sp_dropuser, sp_helpdb, CREATE DATABASE sp_changegroup ──────────────────────────────────────────────────────────────────────────── Function Changes a user's group. Syntax sp_changegroup grpname, username Examples sp_changegroup fort_mudge, albert The user albert is now a member of the fort_mudge group. It doesn't matter what group albert belonged to before. Options grpname The name of the group. The group must already exist in the current database. username The name of the user to add to the group. The user must already exist in the current database. Comments Executing sp_changegroup adds the specified user to the specified group. The user is dropped from the group he or she belongs to and is added to the one specified by grpname. New database users can be added to groups at the same time they are given access to the database with sp_adduser. Every user is a member of the default group, public, if he or she has not been explicitly added to some other group with sp_adduser or sp_changegroup. A user can be a member of only one group. Groups are used as a collective name for granting and revoking permissions. If you use public as the grpname parameter, it must be enclosed in quotation marks because it is an SQL keyword. Messages Group changed. The user now belongs to the specified group. No such group exists. The specified group doesn't exist in the current database. No such user exists. The specified user doesn't exist in the current database. Permissions Execute permission defaults to the Database Owner. Tables Used sysusers See Also sp_addgroup, sp_adduser, sp_dropgroup, sp_helpgroup, GRANT, REVOKE sp_commonkey ──────────────────────────────────────────────────────────────────────────── Function Defines a common key between two tables or views. Syntax sp_commonkey tabaname, tabbname, col1a, col1b [, col2a, col2b, ..., col8a, col8b] Examples sp_commonkey projects, departments, empid, empid Assume two tables, projects and departments, each with a column named empid. This statement defines a frequently used join on the two columns. Options tabaname The name of the first table or view to be joined. tabbname The name of the second table or view to be joined. col1a The name of the first column in the tabaname table or view that makes up the common key. At least one pair of columns (one column from the first table or view and one from the second table or view) must be specified. col1b The name of the partner column in the tabbname table or view that is joined with col1a in the tabaname table or view. Comments Common keys are created to make explicit a logical relationship that is implicit in your database design. The information can be used by an application. Executing sp_commonkey adds the key to the syskeys system table. To display a report on the common keys that have been defined, execute sp_helpkey. You must be the owner of at least one of the two tables or views to define a common key between them. The number of columns from the first table or view must be the same as the number of columns from the second table or view. Up to eight columns from each table or view can participate in the common key. The datatypes of the common columns must agree, but the length specification, if applicable, and nulltypes need not agree. Common keys have been assigned within the master and model databases. Messages New common key added. The common key between the specified tables or views has been added to syskeys. First table in the common key doesn't exist. The table or view you gave as tabaname doesn't exist in the current database. Second table in the common key doesn't exist. The table or view you gave as tabbname doesn't exist in the current database. Only the table owner may define its common key. You aren't the owner of the tabaname table or view. The tables or views have no such first column or the columns are of different types. Two cases generate this message: either the column pair that you specified doesn't exist or the columns in the pair are different types. Permissions Execute permission defaults to the owner of tabaname or tabbname. Tables Used syscolumns, syskeys, sysobjects See Also sp_dropkey, sp_foreignkey, sp_helpjoins, sp_helpkey, sp_primarykey, Joins sp_configure ──────────────────────────────────────────────────────────────────────────── Function Displays or changes configuration options. Syntax sp_configure [configname [, configvalue]] Examples A. sp_configure Example A displays a list of all the configuration options with their current and permissible range of values, similar to the following: ╓┌─┌───────────────────┌─────────┌─────────┌─────────────┌───────────────────╖ ──────────────────────────────────────────────────────────────────────────── name minimum maximum config_value run_value ----------------- -------- -------- ------------ --------- recovery interval 1 32767 4 3 allow updates 0 1 0 0 user connections* 5 35 0 25 memory 1000 14000 0 1700 open databases 5 100 0 10 locks 5000 50000 0 5000 open objects 100 10000 0 500 procedure cache 1 99 0 20 fill factor 0 100 0 0 time slice 50 1000 0 100 database size 2 10000 0 2 media retention 0 365 0 0 recovery flags 0 1 0 0 ──────────────────────────────────────────────────────────────────────────── recovery flags 0 1 0 0 serial number 1 999999 0 0 (14 rows affected) * SQL Server maximum is 250; the effective number is based upon your system configuration. The maximum number of connections in your computer environment is stored in the @MAX_CONNECTIONS global variable. The config_value column of the report contains the value to which the configuration option has been set with sp_configure. It changes after you execute sp_configure. (This is the value in sysconfigures.value.) The run_value column contains the value SQL Server is using. It changes after you execute the RECONFIGURE statement. (This is the value in syscurconfigs.value.) B. sp_configure "recovery interval", 3 Example B sets the system recovery interval to 3 minutes. Options configname The name of the configuration option. SQL Server understands any unique string that is part of the configuration name. configvalue The value for the configuration option. Comments All users can execute sp_configure with no parameters (SQL Server displays a list of all the configuration options and their current values) or with one parameter, the name of the configuration options (SQL Server displays the value for that variable). The System Administrator can execute sp_configure with both parameters to change the value of one of the configuration options. The second parameter is the new config_value. After sp_configure has been executed, the System Administrator must execute the RECONFIGURE statement to install the changed value. Then, for all variables except allow updates and recovery interval, the System Administrator must restart SQL Server. Use RECONFIGURE WITH OVERRIDE when you set allow updates on or when you set a configuration option to a value that SQL Server considers less than optimal. To instruct SQL Server to supply a default configuration option, give the value 0 as the config_value. See the SQL Server System Administrator's Guide for details. Messages Configuration option changed. Run the RECONFIGURE statement to install. After changing a configuration option with sp_configure, the change does not take effect until the RECONFIGURE statement is executed and (for all but allow updates and recovery interval) SQL Server is restarted. Configuration option doesn't exist. The name supplied as the config_name parameter is unknown. Configuration option is not unique. The name supplied as the config_name parameter is not unique. No configuration option was changed. For example, two of the configuration options are recovery interval and recovery flags. Using recovery for the config_name parameter generates this message because it matches both names. The complete names that match the string supplied are printed out so that you can see how to make the config_name more specific. Only the System Administrator may change configuration parameters. Although any user can examine the state of the configuration options, only the System Administrator can change them. Configuration option value is not legal. The config_value supplied is not in the range of permissible values for the specified configuration option. For a display of the range of permissible values, rerun sp_configure with the name of the configuration option as the only parameter. A config_value of 0 is always legal. It instructs SQL Server to set the configuration value to its default. Permissions Execute permission for no parameters or the first parameter defaults to all users. Execute permission for both parameters defaults to the System Administrator. Tables Used sysconfigures, syscurconfigs, spt_values See Also sp_dboption, RECONFIGURE sp_dboption ──────────────────────────────────────────────────────────────────────────── Function Displays or changes database options. Syntax sp_dboption [dbname, optname, {true | false}] Examples A. sp_dboption Example A displays a list of the database options: Settable database options. database_options ----------------------- ALL SETTABLE OPTIONS dbo use only no chkpt on recovery read only select into/bulkcopy single user trunc. log on chkpt. (7 rows affected) B. use master go sp_dboption pubs, read, true use pubs go checkpoint Example B makes the pubs database read only. C. use master go sp_dboption pubs, read, false use pubs go checkpoint Example C makes the pubs database writable again. Options dbname The name of the database in which you want to set the option. You must be using master to execute sp_dboption with parameters (that is, to change a database option). However, the database name cannot be master; you cannot change the master database option settings. optname The name of the option you want to set or unset. SQL Server understands any unique string that is part of the option name. Use quotation marks around the option name if it includes embedded blanks. true, false Use true if you want to set the option and false if you want to turn off the option. Comments You cannot change any of the database option settings for the master database. To display a list of the database options the user can set, execute sp_dboption with no parameters from inside the master database. For a report on which database options are set in a particular database, execute sp_helpdb. The Database Owner or System Administrator can set or unset particular database options for all new databases by executing sp_dboption on the model database. After sp_dboption has been executed, the change does not take effect until the CHECKPOINT statement is executed in the database for which the option was changed. If you change a database option with sp_dboption inside a user-defined transaction and then roll back that transaction, you must execute a CHECKPOINT statement for the change to take effect. Here's an example: begin tran use master go sp_dboption orderentry, single, true go use orderentry go checkpoint go rollback tran go /* ** If the following CHECKPOINT is not executed, ** the orderentry ** database remains single-user. */ checkpoint go While the dbo use only option is set (or true), only the Database Owner can use the database. The read only option means that users can retrieve data from the database but can't modify anything. When single user is set to true, only one user at a time can access the database. The trunc. log on chkpt. option means that the transaction log is truncated (committed transactions are removed) every time the CHECKPOINT checking process occurs (usually more than once per minute). It may be useful to turn this option on while doing development work, to prevent the log from growing. While the trunc. log on chkpt. option is set, you cannot use DUMP TRANSACTION, since dumps from the truncated transaction log dumps cannot be used to recover from a media failure. Issuing the DUMP TRANSACTION statement produces an error message instructing you to use DUMP DATABASE instead. The select into/bulkcopy option must be set if you want to use SELECT INTO on a permanent table or do a "fast" bulk copy into a table that has no indexes. The fast version of bulk copy (using bcp) copies data into tables that have no indexes. Since these operations are not logged, DUMP TRANSACTION is prohibited when this option is set. The user is instructed to use DUMP DATABASE instead. You do not have to set the select into/bulkcopy option to use SELECT INTO on a temporary table since the temporary database is never recovered. The option need not be set to run bcp on a table that has indexes because tables with indexes are always copied with the slower version and are logged. By default, the select into/bulkcopy option is off in newly created databases. To change the default situation, set this option in the model database. The no chkpt on recovery option is set (true) when an up-to-date copy of a database is kept. In these situations, there is a "primary" and a "secondary" database. Initially, the primary database is dumped and loaded into the secondary database. Then, at intervals, the transaction log of the primary database is dumped and loaded into the secondary database. If this option is off (false), the default condition, a checkpoint record is added to a database after it is recovered due to restarting SQL Server. This checkpoint, which insures that the recovery mechanism won't unnecessarily be rerun, changes the sequence number and causes a subsequent load of the transaction log from the primary database to fail. If this option is on (true), no checkpoint record is added to a database after it is recovered, so that subsequent transaction log dumps from the primary database can be loaded into it. See the SQL Server System Administrator's Guide for additional details on database options. Messages No such database -- run sp_helpdb to list databases. No database with the supplied name exists. Run sp_helpdb to get a list of databases. The `master' database's options cannot be changed. No one can change any of master database option settings. Settable database options. Executing sp_dboption with no parameters displays a list of the options the user can set. Database option doesn't exist or can't be set by user. Run sp_dboption with no parameters to see options. This message is generated in two cases: either the option doesn't exist or the user does not have permission to set or unset it. Run the system procedure with no parameters to display a list of the options the user can set. Database option is not unique. The name supplied as the optname parameter is not unique. No database option value was changed. For example, two of the database options are dbo use only and read only. Using only for the optname parameter generates this message because it matches both names. The complete names that match the supplied string are printed out so that you can see how to make the optname more specific. Only the System Administrator or the owner of the database may set db options. Although all users can look at a list of the settable database options, only the System Administrator or Database Owner can set or unset them. You aren't the System Administrator or the Database Owner. You must be in the `master' database to change database options. To change a database option (of any database other than master), execute the sp_dboption system procedure with the appropriate parameters, while using master. usage: sp_dboption [dbname, optname, {true | false}] Either the optname parameter was omitted or the optvalue parameter was something other than true or false. Run the CHECKPOINT command in the database that was changed. The change in the database option takes effect after the CHECKPOINT statement is run. Permissions Execute permission for no parameters (display options only) defaults to all users. Execute permission for parameters (change an option) defaults to the System Administrator and Database Owner of the database for which the option is to be changed. Tables Used master.dbo.spt_values, master.dbo.sysdatabases See Also sp_configure, sp_helpdb, sp_helpjoins, CHECKPOINT, SELECT sp_defaultdb ──────────────────────────────────────────────────────────────────────────── Function Changes a user's default database. Syntax sp_defaultdb login_id, defdb Examples sp_defaultdb victoria, pubs This example sets the default database for victoria to pubs. Options login_id The login ID of the user. defdb The name of the user's default, or "home," database (the database to which the user is connected when he or she logs in). Comments A default database can be set either with sp_defaultdb or with sp_addlogin when the user's login ID is first added to SQL Server. The database to which users are connected if no default database has been specified, either with sp_defaultdb or sp_addlogin, is master. After sp_defaultdb is executed, the user is connected to the new defdb the next time he or she logs in. However, this system procedure does not automatically give the user access to that database. The Database Owner must give the user database access through sp_adduser or sp_addalias, or there must be a guest user in the database's sysusers table. If the user does not have access to the database by any of these means, she or he is connected to master, and an error message is displayed. If a user's default database is dropped, the user is connected to master on his or her next login, and an error message is displayed. The System Administrator can change anyone's default database with this system procedure. All others can change only their own default database. Messages Default database changed. The default database for login_id has been changed successfully. No such account exists. There is no login ID for login_id. Nothing was changed. Database name not valid -- default not changed. There is no database with the specified name. You can't change someone else's default database. Every user can change his or her own default database, but only the System Administrator can change other users' default databases. You aren't the System Administrator. Permissions Execute permission defaults to public for his or her own login ID only. Execute permission defaults to the System Administrator for any login ID. Tables Used master.dbo.sysdatabases, master.dbo.syslogins See Also sp_addlogin, sp_droplogin, sp_password, USE sp_depends ──────────────────────────────────────────────────────────────────────────── Function Displays information about database object dependencies─the views and procedures that depend on the table or view specified, and the tables and views that are depended on by the view or procedure specified. Syntax sp_depends objname Examples A. sp_depends sysobjects Example A lists the database objects that depend on the sysobjects table. B. sp_depends titleview Things the object references in the current database. object type updated selected -------------------- ---------------- ------- -------- dbo.authors user table no no dbo.titleauthor user table no no dbo.titles user table no no Things inside the current database that reference the object. object type ------------------ --------------- dbo.tview2 view Options objname The name of the database object that you want to examine for dependencies. It can be a table, view, stored procedure, or trigger. Comments Executing sp_depends lists all the objects, if any, that depend on objname and all the objects, if any, that objname depends on. For example, views depend on one or more tables and can have procedures or other views that depend on them. The sp_depends system procedure determines the dependencies by looking at sysdepends. An object that references another object is considered dependent on that object. The updated and selected columns in the report from sp_depends are meaningful if the object being reported on is a stored procedure or trigger. The values in these columns indicate whether the stored procedure or trigger updates or selects from that object. References to objects outside the current database are not reported. Messages No such object in the current database. The object name supplied for the objname parameter does not exist in the current database. Things the object references in the current database. These are the objects in the current database that objname depends on. Things inside the current database that reference the object. These are the objects in the current database that reference objname. Object doesn't reference any object and no objects reference it. Nothing depends upon objname, and objname doesn't reference any objects. Permissions Execute permission defaults to the public group. Tables Used master.dbo.spt_values, master.dbo.sysdatabases, sysdepends, sysobjects See Also sp_help, CREATE PROCEDURE, CREATE TABLE, CREATE VIEW, EXECUTE, Views sp_diskdefault ──────────────────────────────────────────────────────────────────────────── Function Sets a database device's status to defaulton or defaultoff. This status indicates whether a database device can be used for database storage when the user does not specify a database device, or specifies DEFAULT with CREATE DATABASE or ALTER DATABASE statements. Syntax sp_diskdefault database_device, {defaulton | defaultoff} Examples sp_diskdefault master, defaultoff The master database device is no longer used by CREATE DATABASE or ALTER DATABASE for default storage of a database. Options database_device The logical name of the database device. It must be a database device rather than a dump device. defaulton, defaultoff Use defaulton if the specified database device is to be designated a default database device. Use defaultoff if the specified database device is not to be designated a default database device. The defaulton option will most frequently be used after a database device has been added to the system with DISK INIT. The defaultoff option will most frequently be used to change the default status of the master database device (which is on when SQL Server is first installed). Comments A default database device is one that is used for database storage by CREATE DATABASE or ALTER DATABASE if the user does not specify a database devicename or if the user gives the DEFAULT keyword. When you first install SQL Server, the master database device is the only default database device. To find out which database devices are default database devices, execute the sp_helpdevice system procedure. Messages No such device exists. The database devicename supplied for the database_device doesn't exist on SQL Server. Run the sp_helpdevice system procedure without a parameter to see a list of all database devices. To add a new database device to the system, use the DISK INIT statement. The device name supplied is not a database disk. The database devicename supplied is in sysdevices, but it is a dump device rather than a database device. Run the sp_helpdevice system procedure without a parameter to see a list of all database devices. To add a new database device to the system, use the DISK INIT statement. Only the System Administrator may execute this procedure. Only the System Administrator can change the default status of database devices. Usage: sp_diskdefault logical_name {defaulton | defaultoff}. The second parameter must be either defaulton or defaultoff. Permissions Execute permission defaults to the System Administrator. Tables Used master.dbo.sysdevices See Also sp_helpdevice, ALTER DATABASE, CREATE DATABASE, DISK INIT sp_dropalias ──────────────────────────────────────────────────────────────────────────── Function Removes the alias username identity that had been established with sp_addalias. Syntax sp_dropalias login_id Examples sp_dropalias victoria Assuming that victoria had an alias (for example, to the Database Owner) in the current database, this statement drops victoria's alias from the database. Options login_id The name (in master.dbo.syslogins) of the user who has an alias. Comments Executing the sp_dropalias system procedure deletes an alternate suid mapping for a user from the sysalternates table. When a user's alias is dropped, he or she no longer has access to the database for which the alias was created. Messages Alias user dropped. The user no longer has an alias in the current database. He or she cannot use the database until the Database Owner reinstates him or her using sp_adduser or sp_addalias. No alias for specified user exists. The named user doesn't have an alias in the current database. No login with the specified name exists. The login_id you supplied has no account on SQL Server. No action was taken. Permissions Execute permission defaults to the Database Owner. Tables Used sysalternates See Also sp_addalias, sp_adduser, sp_changedbowner, sp_droplogin, sp_dropuser, sp_helpuser, USE sp_dropdevice ──────────────────────────────────────────────────────────────────────────── Function Drops an SQL Server database device or dump device. Syntax sp_dropdevice logical_name Examples A. sp_dropdevice disk5 Example A drops the disk5 dump device from SQL Server. B. sp_dropdevice fredsdata Example B drops the fredsdata database device from SQL Server. This database device must not be in use by any databases. Options logical_name The logical name of the database device or dump device as listed in master.dbo.sysdevices.name. Comments The sp_dropdevice system procedure drops a database device or dump device from SQL Server, deleting the entry from master.dbo.sysdevices. ──────────────────────────────────────────────────────────────────────────── WARNING After executing sp_dropdevice, you must restart SQL Server because the kernel has a process that is accessing the dropped database device; there is no way to kill the process. Failure to restart SQL Server can cause errors. Also, restarting SQL Server frees up the database device or dump device. ──────────────────────────────────────────────────────────────────────────── Messages Device dropped. The database device or dump device has been dropped from the master.dbo.sysdevices table. No device with specified logical name exists. You tried to drop a database device or dump device that doesn't exist on SQL Server. Only the System Administrator may execute this procedure. Only the System Administrator can add or drop database devices or dump devices. Device is being used by a database. You can't drop it. Only database devices that are not in use can be dropped. You must drop all the databases associated with the database device before dropping the database device. Permissions Execute permission defaults to the System Administrator. Tables Used master.dbo.sysdevices See Also sp_addumpdevice, sp_helpdb, sp_helpdevice, DROP DATABASE sp_dropgroup ──────────────────────────────────────────────────────────────────────────── Function Drops a group from a database. Syntax sp_dropgroup grpname Examples sp_changegroup accounting, martha sp_changegroup "public", george sp_dropgroup purchasing The purchasing group has merged with the accounting group. Members of the purchasing group, martha and george, are moved to the accounting group and the public group, respectively. Then the purchasing group is dropped. The groupname "public" is in quotation marks because public is a reserved word. Options grpname The name of a group in the current database. Comments Executing sp_dropgroup drops a groupname from a database's sysusers table. You can't drop a group if it has members. You must execute sp_changegroup for each member before you can drop the group. Messages Group has been dropped. The group no longer exists in the current database. No group with the specified name exists. The specified group doesn't exist. Can't drop the group 'public'. The public group exists in every database. It is the group that all users belong to by default and cannot be dropped. Group has members. It must be empty before it can be dropped. Groups with members can't be dropped. Reassign the members of the group to another group using sp_changegroup. When this error message is generated, a list of the group members is displayed. Permissions Execute permission defaults to the Database Owner. Tables Used sysobjects, sysusers See Also sp_addgroup, sp_adduser, sp_changegroup, sp_dropuser, sp_helpgroup, GRANT, REVOKE, USE sp_dropkey ──────────────────────────────────────────────────────────────────────────── Function Removes a key from the syskeys table that had been defined using sp_primarykey, sp_foreignkey, or sp_commonkey. Syntax sp_dropkey keytype, tabname [, deptabname] Examples A. sp_dropkey primary, employees Example A drops the primary key for the employees table. Any foreign keys that were dependent on the primary key on employees are also dropped. B. sp_dropkey common, employees, projects Example B drops the common keys between the employees and projects tables. Options keytype The type of key to be dropped. The keytype must be primary, foreign, or common. tabname The name of the table that contains the key to be dropped. deptabname The name of the dependent table in the key to be dropped. Use this option only if the keytype is foreign. The deptabname is the table that contains the foreign key. If the keytype is primary, this parameter is ignored since primary keys have no dependent table. Comments Executing sp_dropkey deletes the specified key from syskeys. Only the owner of a table can drop a key on that table. Dropping a primary key automatically drops any foreign keys associated with it, but dropping a foreign key has no effect on a primary key specified on that table. Keys are created to make explicit a logical relationship that is implicit in your database design. This information can be used by an application program. Executing sp_commonkey, sp_primarykey, or sp_foreignkey adds the key to the syskeys system table. To display a report on the keys that have been defined, execute sp_helpkey. Messages Usage: sp_dropkey {primary | foreign | common}, tabname [,deptabname]. Type must be 'primary', 'foreign', or 'common'. The keytype should specify the type of key to drop. Table or view name must be in 'current' database and owned by you. You can't drop keys on tables in other databases. The table named doesn't exist in the current database. The tabname supplied isn't a table or view in the current database. You must be the owner of the table to drop its key. You aren't the owner of tabname, so you can't drop the key. No primary key for the table exists. The tabname has no primary key defined. Primary key for the table dropped. The primary key was successfully dropped. It has been deleted from syskeys. Dependent foreign keys were also dropped. When a primary key is dropped, any foreign keys that depend on it are also dropped. You need to supply the dependent table as the third parameter. When dropping a foreign or common key, both the tabname and deptabname tables must be named. The dependent table doesn't exist in the current database. The name supplied for the deptabname parameter isn't a table or view in the current database. No foreign key for the table exists. The tabname has no foreign key defined. Foreign key dropped. The foreign key has been successfully dropped. It has been deleted from syskeys. No common keys exist between the two tables supplied. There are no common keys between the tabname and deptabname tables. No action was taken. Common keys dropped. The common keys were successfully dropped. They have been deleted from syskeys. Permissions Execute permission defaults to the owner of tabname. Tables Used syskeys, sysobjects See Also sp_commonkey, sp_foreignkey, sp_helpkey, sp_primarykey sp_droplogin ──────────────────────────────────────────────────────────────────────────── Function Drops an SQL Server user by deleting the user's entry in master.dbo.syslogins. Syntax sp_droplogin login_id Examples sp_droplogin victoria The user victoria has been dropped from SQL Server. Options login_id The name of the user as listed in master.dbo.syslogins. Comments Executing sp_droplogin drops a user login ID from SQL Server, deleting the user's entry from master.dbo.syslogins. The sp_droplogin procedure checks the current database (but no others) to see if the login ID to be dropped is a user there. You cannot execute sp_droplogin if the user exists in the current database. Messages Login dropped. The user's entry in master.dbo.syslogins has been deleted. The user no longer has access to SQL Server. No login exists with supplied name. The specified login ID does not exist. Only the System Administrator may execute this procedure. Only the System Administrator can add or drop login IDs. You aren't the System Administrator. User exists in current database. Drop user before dropping login. The specified login ID is a user in the current database. Drop the user from the current database with the sp_dropuser procedure. When this error message is generated, the database name, user's login ID, and user's name in the database are displayed. Permissions Execute permission defaults to the System Administrator. Tables Used master.dbo.syslogins, sysusers See Also sp_addlogin, sp_changedbowner, sp_dropuser, sp_helpuser sp_droptype ──────────────────────────────────────────────────────────────────────────── Function Drops a user-defined datatype by deleting the type from systypes. Syntax sp_droptype typename Examples sp_droptype birthday This example drops the user-defined datatype named birthday. Options typename The name of a user-defined datatype that you own. Comments Executing sp_droptype deletes a user-defined datatype from systypes. A user datatype cannot be dropped if tables or other database objects reference it. Messages Type has been dropped. The user-defined type no longer exists in the current database. The type doesn't exist or you don't own it. You do not own a user datatype with that name. Type is being used. You cannot drop it. A user datatype referenced by a table or other database object cannot be dropped. Drop the tables and/or procedures first. Permissions Execute permission defaults to the Database Owner and datatype owner. Tables Used syscolumns, sysobjects, systypes, sysusers See Also sp_addtype, sp_rename sp_dropuser ──────────────────────────────────────────────────────────────────────────── Function Drops a user from the current database by deleting the entry from sysusers. Syntax sp_dropuser username Examples sp_dropuser albert This statement drops the user albert from the current database. The user albert may no longer use the database. Options username The user's name in the current database's sysusers table. Comments Executing sp_dropuser deletes a user from the current database by deleting the user's row from sysusers. ──────────────────────────────────────────────────────────────────────────── WARNING Do not drop users who own database objects. Before dropping a user, first check to see if that user owns any database objects. If the user does own database objects, drop the objects or change their ownership before dropping the user. ──────────────────────────────────────────────────────────────────────────── The Database Owner of a database cannot be dropped. If other users have an alias to the user being dropped, their aliases are dropped. They are no longer able to access the database. Messages User has been dropped from current database. The specified user is no longer known to the database. Dependent aliases were also dropped. Other users had an alias to the user being dropped. Their aliases have been dropped, and they can no longer access the database. No user exists with the supplied name. The specified user doesn't exist in the current database. You cannot drop the 'Database Owner'. The Database Owner cannot be dropped, even by the System Administrator. You cannot drop user because he or she owns objects in database. Users who own objects in the current database cannot be dropped. Drop the owned objects first. When this error message is generated, a list of objects and their owners is displayed. You cannot drop user because he or she owns types in database. Users who own user datatypes in the current database cannot be dropped. Drop the owned datatypes first. When this error message is generated, a list of datatypes and their owners is displayed. Permissions Execute permission defaults to the Database Owner. Tables Used sysalternates, sysobjects, systypes, sysusers See Also sp_adduser, sp_droplogin, GRANT, REVOKE, USE sp_foreignkey ──────────────────────────────────────────────────────────────────────────── Function Defines a foreign key on a table or view. Syntax sp_foreignkey tabname, pktabname, col1 [, col2, col3, ..., col8] Examples A. sp_foreignkey titles, publishers, pub_id The primary key of the publishers table is the pub_id column. The titles table also contains a pub_id column, which is a foreign key of publishers. B. sp_foreignkey orders, parts, part, subpart The primary key of the parts table has been defined with sp_primarykey as the partnumber and subpartnumber columns. The orders table contains the columns part and subpart, which make up a foreign key of parts. Options tabname The name of the table or view that contains the foreign key. pktabname The name of the table or view that has the primary key to which the foreign key applies. col1 The name of the first column that makes up the foreign key. The foreign key must consist of at least one column and can have a maximum of eight columns. Comments Executing sp_foreignkey adds the key to the syskeys table. Keys are created to make explicit a logical relationship that is implicit in your database design. To display a report on the keys that have been defined, execute sp_helpkey. You must be the owner of the table or view to define its foreign key. The primary key of pktabname table or view must already be defined. The number and order of columns that make up the foreign key must be the same as the number and order of columns that make up the primary key. The datatypes of the primary and foreign keys must agree, but the lengths and nulltypes need not agree. When you define a foreign key, a common key is automatically defined between the pktabname and tabname tables or views, and added to syskeys. Foreign keys have been assigned within the master and model databases. Messages Foreign key table doesn't exist. The table or view specified with the tabname parameter doesn't exist in the current database. Primary key table doesn't exist. The table or view specified with the pktabname parameter doesn't exist in the current database or doesn't have a primary key defined. Only the owner of the table may define a foreign key. You are not the owner of the table or view. The table has no such first column. The table or view specified with the tabname parameter does not have a column of the specified name. Primary key does not exist with the same number of columns as the foreign key. The number of columns in the foreign key of tabname must be the same as the number of columns in the primary key of pktabname. Datatypes of the first column in the keys are different. The datatypes of the columns of the foreign key of tabname and the primary key of pktabname must be the same. New foreign key and common join added. The foreign key has been defined and added to syskeys. A common key has also been added to syskeys. Permissions Execute permission defaults to the tabname owner. Tables Used syscolumns, syskeys, sysobjects See Also sp_commonkey, sp_dropkey, sp_helpjoins, sp_helpkey, sp_primarykey, CREATE TRIGGER sp_help ──────────────────────────────────────────────────────────────────────────── Function Reports information about a database object (any object listed in sysobjects) or about an SQL-Server-supplied or user-defined datatype. Syntax sp_help [objname] Examples A. sp_help Example A displays a brief listing of each object in sysobjects, giving its name, owner, and object type, and of each user-defined datatype in systypes, giving its name, storage type, case sensitivity, length, nulltype, default name, and rule name. The nulltype is 0 if null values are not allowed and 1 if null values are allowed. B. sp_help publishers Example B displays information about the publishers table: Name Owner Type ----------- ----------- - ------------ publishers dbo user table (0 rows affected) Data_located_on_segment When_created ----------------------- -------------------- default Sep 15 1987 2:03PM (0 rows affected) ╓┌───────┌────────────┌────────────────┌───────┌────────────┌─────────────┌── ───────────────────────────────────────────────────────────────────────────── ───────────────────────────────────────────────────────────────────────────── Column_name Type Length Null values Rule Default_name ----------- --------------- ------ ----------- --- ------------ pub_id char nocase 4 0 NULL pub_ pub_name varchar nocase 40 1 NULL NULL city varchar nocase 20 1 NULL NULL state char nocase 2 1 NULL NULL (9 rows affected) index_name index_description index_keys ---------------- -------------------- ----------- pubind clustered, unique pub_id (1 row affected) No defined keys for this object. Options objname The name of any object in sysobjects or any user-defined datatype in systypes. Database names are not acceptable. Comments The procedure looks for an object in the current database only. System procedures do not work on temporary tables because a procedure cannot access a temporary object unless it created the object itself during the current session. If sp_help prints "nocase" after char and varchar datatypes, the server is not case sensitive. Messages Object must be in your current database. The procedure gives information about objects in the current database only. You have included the database name in the objname parameter. Object does not exist in this database. The specified object does not exist in the current database. No defined keys for this object. The table or view has no defined keys. Tables Used master.dbo.spt_values, syscolumns, sysobjects, systypes Permissions Execute permission defaults to the public group. See Also sp_helpgroup, sp_helpindex, sp_helprotect, sp_helpuser sp_helpdb ──────────────────────────────────────────────────────────────────────────── Function Reports information about a database or about all databases. Syntax sp_helpdb [dbname] Examples A. sp_helpdb pubs Example A displays information about the pubs database: name db_size owner dbid created status ------ -------- -------- ------ ----------- ---------------- pubs 2 MB sa 4 Sep 15 1987 no options set (1 row affected) device size usage --------------------- ------------ ---------------- master 2 MB data and log (1 row affected) B. sp_helpdb Example B displays information about all the databases on SQL Server. Options dbname The name of the database you want information on. Omitting this optional parameter results in a report on all databases. Comments Executing sp_helpdb reports on the specified database when the dbname is supplied or on all databases in master.dbo.sysdatabases when no parameter is supplied. Messages No such database exists. The specified database doesn't exist on SQL Server. Run the procedure without the dbname parameter to see a list of all the databases. Permissions Execute permission defaults to the public group. Tables Used #spdbdesc, master.dbo.spt_values, master.dbo.sysdatabases, master.dbo.sysdevices, master.dbo.syslogins, master.dbo.sysusages See Also sp_configure, sp_dboption, sp_renamedb, ALTER DATABASE, CREATE DATABASE sp_helpdevice ──────────────────────────────────────────────────────────────────────────── Function Reports information about SQL Server's database devices and dump devices. Syntax sp_helpdevice [logical_name] Examples A. sp_helpdevice device_name physical_name description status cntrltype device_number low high --------------- ----------------------------- ------------------------------ ----------- ------------- ------------- ----- ------ diskdump nul disk dump device 16 2 0 0 20000 diskettedumpa a:sqltable.dat diskette, 1.2 Mb, dump device 16 3 0 0 19 diskettedumpb b:sqltable.dat diskette, 1.2 Mb, dump device 16 4 0 0 19 master c:\sql\data\master.dat special, default disk, physical disk, 10 Mb 3 0 0 0 5119 (4 rows affected) Example A reports information about all database devices and dump devices on SQL Server. B. sp_helpdevice diskdump Example B reports information about the device named diskdump. Options logical_name The logical name of the database device or dump device you want information on. If you omit this parameter, information on all the database files and dump devices is displayed. Comments Executing sp_helpdevice displays information on the specified database device or dump device when logical_name is given or on all database devices and dump devices in master.dbo.sysdevices when no parameter is given. The sysdevices table contains database devices and dump devices. Database devices can be designated default database devices, which means that they can be used for database storage when a user executes the CREATE DATABASE or ALTER DATABASE statement and does not specify a database devicename, or gives the keyword DEFAULT. To make a database device a default database device, execute the sp_diskdefault system procedure. Database devices are added to the system with the DISK INIT statement. Dump devices are added to the system with the sp_addumpdevice system procedure. The status column in the report from sp_helpdevice contains a number that corresponds to the status description in the description column. The cntrltype column specifies the controller number of the device. For hard disk dump devices, the controller number is 2; for diskette dump devices, the controller number is 3 or 4. For database devices, it is almost always 0 (unless your installation has a special type of disk controller). The device_number column is 0 for dump devices, 0 for the master database device, and between 1 and 9 for other database devices. The low and high columns represent virtual page numbers for database devices devices and disk dump devices. The high and low columns represent media capacity for diskette dump devices (high - low = number of 62K blocks). Messages No such i/o device exists. The database device or dump device name supplied for the logical_name parameter doesn't exist on SQL Server. Run the procedure without the logical_name parameter to see a list of all devices. Permissions Execute permission defaults to the public group. Tables Used #spdevtab, master.dbo.spt_values, master.dbo.sysdevices See Also sp_addumpdevice, sp_configure, sp_diskdefault, sp_dropdevice, sp_helpdb, sp_logdevice, sp_who, DISK INIT, DUMP DATABASE, DUMP TRANSACTION, LOAD DATABASE, LOAD TRANSACTION sp_helpgroup ──────────────────────────────────────────────────────────────────────────── Function Reports information on a group or on all groups in the current database. Syntax sp_helpgroup [grpname] Examples A. sp_helpgroup hackers Example A displays information about the hackers group: Group_name Group_id Users_in_group Userid -------------- ---------- ------------- ------ hackers 16384 ann 4 hackers 16384 judy 3 (3 rows affected) B. sp_helpgroup Example B displays information about all the groups in the current database: ──────────────────────────────────────────────────────────────────────────── Group_name Group_id ------------ -------- hackers 16384 public 0 (2 rows affected) Options grpname The name of a group in the database that has been created with sp_addgroup. Comments Executing sp_helpgroup reports on the specified group or, if no parameter is supplied, on all groups in the database. Messages No such group exists in the current database. The specified group does not exist in the current database. Execute the procedure without the grpname parameter to see a list of all the groups in the database. Permissions Execute permission defaults to the public group. Tables Used sysusers See Also sp_addgroup, sp_changegroup, sp_dropgroup, sp_helprotect, sp_helpuser, GRANT, REVOKE sp_helpindex ──────────────────────────────────────────────────────────────────────────── Function Reports index information on a table. Syntax sp_helpindex tabname Examples sp_helpindex sysobjects This statement reports on the types of indexes on the sysobjects table: index_name index_description index_keys -------------------- -------------------- --------------- sysobjects clustered, unique id ncsysobjects nonclustered, unique name, uid (2 rows affected) Options tabname The name of a table in the current database. Comments Executing sp_helpindex lists up to eight indexes on a table. ──────────────────────────────────────────────────────────────────────────── WARNING Only eight indexes are displayed when you execute sp_helpindex. The sysindexes system table contains information on all indexes. ──────────────────────────────────────────────────────────────────────────── Messages Table does not exist. The name you gave for the tabname parameter does not exist in the current database. Table does not have any indexes. The table you named has no indexes. Table name must be in 'current' database. The name you gave for the tabname parameter includes a database reference. Name references must be local to the current database. Permissions Execute permission defaults to the public group. Tables Used #spindtab, master.dbo.spt_values, sysindexes See Also sp_help, sp_helpkey, CREATE INDEX, DROP INDEX, UPDATE STATISTICS sp_helpjoins ──────────────────────────────────────────────────────────────────────────── Function Lists the columns in two tables or views that are likely to be joined. Syntax sp_helpjoins lefttab, righttab Examples sp_helpjoins sysobjects, syscolumns This statement displays a list of columns that are likely to be joined in the sysobjects and syscolumns tables: a1 a2 b1 b2 c1 c2 d1 d2 e1 -------- -------- -------- -------- -------- ------- -------- -------- -------- -------- -------- ------- -------- -------- -------- -------- -------- ------- id id NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL NULL (1 row affected) Options lefttab The first table or view. righttab The second table or view. The order of the parameters does not matter. Comments The column pairs that sp_helpjoins displays come from two sources. First, sp_helpjoins checks the syskeys table in the current database to see if any common keys have been defined on the two tables with sp_commonkey. If it doesn't find any common keys there, the procedure applies less restrictive criteria to come up with any keys that may be reasonably joined: it checks for keys with the same user datatypes and if that fails, for columns with the same name and datatype. The sp_helpjoins procedure does not create any joins. Messages First table doesn't exist. The table specified as the lefttab parameter is not a table or view in the current database. Second table doesn't exist. The table specified as the righttab parameter is not a table or view in the current database. Permissions Execute permission defaults to the public group. Tables Used syscolumns, syskeys, sysobjects See Also sp_commonkey, sp_foreignkey, sp_help, sp_helpkey, sp_primarykey, Joins sp_helpkey ──────────────────────────────────────────────────────────────────────────── Function Reports information on primary, foreign, and common keys. Syntax sp_helpkey [objname] Examples sp_helpkey This statement displays information on the keys that have been defined in the database: keytype object related_object object_keys related_keys ------- ------- -------------- ---------- ----------------- primary authors none au_id, *, * *, *, *, *, *, * foreign titleauthor authors au_id, *, * au_id, *, *, *, * (2 rows affected) Options objname The name of a table or view in the current database. If you don't specify a name, the procedure reports on all keys that have been defined in the current database. Comments Executing sp_helpkey lists information about all the primary, foreign, or common key definitions that reference the objname table, or about all the keys in the database. The object_keys and related_keys columns in the report refer to the names of the columns that make up the key. Keys are created to make explicit a logical relationship that is implicit in your database design. The information can be used by an application program. Messages Table or view name must be in 'current' database. The name supplied for the objname parameter included a database reference. You can get key information about local tables or views only. No such table or view in the current database. The name supplied for the objname parameter is not a table or view in the current database. Permissions Execute permission defaults to the public group. Tables Used master.dbo.spt_values, syskeys See Also sp_commonkey, sp_foreignkey, sp_help, sp_primarykey, CREATE TRIGGER sp_helprotect ──────────────────────────────────────────────────────────────────────────── Function Reports permissions by database object and, optionally, by user for that object. Syntax sp_helprotect name [, username] Examples A. grant select on titles to judy grant update on titles to judy revoke update on titles(price) from judy sp_helprotect titles After the series of GRANT and REVOKE statements in example A, executing sp_helprotect titles results in this display: type action user column --------- ----------- ------- ---------- Grant Select public All Grant Select guest All Grant Update guest All Grant Select judy All Grant Update judy title_id Grant Update judy title Grant Update judy type Grant Update judy pub_id Grant Update judy advance Grant Update judy royalty Grant Update judy notes Revoke Update judy price (12 rows affected) B. sp_helprotect judy Example B displays all the permissions that judy has in the database. Options name The name of a table, view, or stored procedure; or the name of a user or group in the current database. username A user's name in the current database. Comments Executing sp_helprotect reports permissions on a database object. If the username parameter is supplied, only those users' permissions on the database object are reported. If name is not an object and username is also supplied, the permissions for the user or group are listed. The procedure looks for objects and users in the current database only. Messages Name must be in 'current' database. The name supplied for the name parameter included a reference to a database. The name must be local to the database. No such user exists in the database. The name supplied for username is not a user or group in the current database. No such object or user exists in the database. The name supplied for the name parameter is not an object, user, or group in the current database. Permissions Execute permission defaults to the public group. Tables Used master.dbo.spt_values, syscolumns, sysobjects, sysprotects, sysusers See Also sp_help, GRANT, REVOKE sp_helpsql ──────────────────────────────────────────────────────────────────────────── Function Provides syntax for SQL statements, system procedures, and other special topics. Syntax sp_helpsql ["topic"] Examples A. sp_helpsql "disk init" Example A displays a brief definition of the DISK INIT statement along with its syntax. B. sp_helpsql functions Example B displays information on SQL Server functions. C. sp_helpsql "cre" Example C displays a list of all statements starting with the letters "CRE" (such as CREATE DATABASE, CREATE DEFAULT, and so on). Options topic The topic on which you want help. Help is available for all SQL statements, all system procedures, and the following special topics: datatypes, expressions, functions, wildcard characters, mathematical functions, and text functions. Be sure to include quotation marks around "topic". Comments You can enter just the first few letters of a topic. If these letters are enough to identify the topic, then help on the topic is displayed. For example, to get help on LOAD DATABASE, you could execute sp_helpsql "load d" To get help on all topics starting with "LOAD," execute sp_helpsql "load" Messages sp_helpsql supplies help on SQL statements, system procedures, and special topics: Syntax: sp_helpsql [topic] You executed sp_helpsql without specifying a topic. Supply a topic and re-execute sp_helpsql. You have entered a non-unique topic. More information can be obtained for the following topics: You have not entered enough letters to identify a unique topic. Supply the complete topic name. For example, enter "DISK INIT" rather than just "DISK". No help available for topic. There is no help available for the topic you entered. Permissions Execute permission defaults to the public group. sp_helptext ──────────────────────────────────────────────────────────────────────────── Function Prints the text of a stored procedure, trigger, view, default, or rule. Syntax sp_helptext objname Examples A. sp_helptext ziprule --------- 1 (1 row affected) text ------------------------------------------------- create rule ziprule as @zip like "[0-9][0-9][0-9][0-9][0-9]" (1 row affected) Example A displays the text of ziprule. Since this rule is in the pubs database, this system procedure must be executed from pubs. B. sp_helptext sp_helptext --------- 3 (1 row affected) text ------------------------------------------------------- /* helptext 1.5 5/4/87 */ create procedure sp_helptext @OBJNAME varchar(92) as /* **, Make sure the @objname is local to the current database. */ if @objname like "%.%.%" and substring(@objname,, 1,, charindex(".",, @objname) - 1) != db_name() begin print "Object must be in your current database." return end /* **, Find out how many lines of text are coming back. */ select count(*) from syscomments where id = object_id(@objname) /* **, Now get the text. */ select text from syscomments where id = object_id(@objname) return (3 rows affected) Example B displays the text of sp_helptext. Since system procedures are stored in master, this system procedure must be executed from master. Options objname The name of the object for which you wish to see the CREATE text. It must be in the current database. Comments Executing sp_helptext prints out the number of rows in syscomments (255 characters long each) that the object occupies, followed by the CREATE text of the object. The system procedure looks for the text in the syscomments table of the current database only. Messages Object must be in your current database. The objname parameter included a database name reference. The objname must be in the current database. Permissions Execute permission defaults to the public group. Tables Used syscomments See Also sp_help, CREATE DEFAULT, CREATE PROCEDURE, CREATE RULE, CREATE TRIGGER, CREATE VIEW sp_helpuser ──────────────────────────────────────────────────────────────────────────── Function Reports information on users of a database. Syntax sp_helpuser [username] Examples A. sp_helpuser Example A displays information about all users in the current database: Users_name ID_in_db Group_name Login_id Default_db ----------- -------- ------------ ---------- ------------ ann 4 hackers ann master dbo 1 public sa master guest 2 public NULL NULL judy 3 hackers judy master (4 rows affected) B. sp_helpuser dbo Example B displays information about the user Database Owner: Users_name ID_in_db Group_name Login_id Default_db ---------- -------- ---------- -------- ------------ dbo 1 public sa master Users aliased to user. Login_id ------------------------------- andy christa howard linda Options username The user's name in the current database. If you don't specify a name, the system procedure reports on all users of the current database. Comments Executing sp_helpuser reports information on a specified user or on all users of the database. If the specified user is not listed in the current database's sysusers table, the procedure checks to see if the user has an alias to another user or is a groupname. Messages The name supplied is a group name. The name specified for the username parameter is a groupname. Users aliased to user. If the user has other users that have an alias to him or her, the names of the other users are listed. The name supplied is aliased to another user. The name supplied is not a user in the database, but has an alias to a user in the database. The name supplied is not a user, group, or aliased. The name supplied is unknown in the database as a login ID, user, or group. Permissions Execute permission defaults to the public group. Tables Used master.dbo.syslogins, sysalternates, sysusers See Also sp_adduser, sp_dropuser, sp_help, sp_helpgroup, GRANT, REVOKE, USE sp_lock ──────────────────────────────────────────────────────────────────────────── Function Reports information on locks. Syntax sp_lock [spid1 [, spid2]] Examples A. sp_lock ╓┌─┌──────────────────┌──────────┌──────────┌──────┌─────────────────────────╖ ──────────────────────────────────────────────────────────────────────────── spid locktype table_id page dbname ----- --------- -------- ----- ----------- 1 Sh_intent 16003088 0 master 4 Ex_extent 0 440 pubs 4 Ex_extent 0 504 pubs 4 Sh_table 112003430 0 pubs 4 Ex_table 240003886 0 pubs (5 rows affected) Example A displays information on all locks currently held in SQL Server. B. sp_lock 1 ──────────────────────────────────────────────────────────────────────────── spid locktype table_id page dbname ----- --------- --------- ----- ------ 1 Sh_intent 16003088 0 master Example B displays information on locks currently held on spid 1. Options spid1 The SQL Server process ID number from master.dbo.sysprocesses. Execute sp_who to get the spid of the lock. The parameter is optional. If it is not supplied, information on all locks is displayed. spid2 Another SQL Server process ID number to check for locks. The parameter is optional. Comments Executing sp_lock reports information on processes that currently hold locks. Use the OBJECT_ID function to get a table's name from its ID number. If the system procedure is executed without parameters, it displays a list of the current locks. The locktype column indicates not only whether the lock is a shared lock (Sh) or an exclusive lock (Ex), but also whether it is held on a table or a page, and whether it is an intent lock, an extent lock, or a demand lock. An intent lock indicates the intention to acquire a shared or exclusive lock on a data page. Setting an intent lock prevents another transaction from acquiring an exclusive lock on the table that contains that page. An extent lock is a lock held on a group of eight database pages while they are being allocated or freed. Extent locks are set while a CREATE or DROP statement is running, or while an INSERT statement that requires new data or index pages is running. A demand lock prevents any more shared locks from being set. It indicates that a transaction is next in line to lock a table or page. Demand locks are necessary because shared locks can overlap so that read transactions keep monopolizing a table or page, forcing a write transaction to wait indefinitely. After waiting on four different read transactions, a write transaction is given a demand lock. As soon as the existing read transactions finish, the write transaction is allowed to proceed. Any new read transactions then have to wait for the write transaction to finish. Permissions Execute permission defaults to the public group. Tables Used master.dbo.spt_values, master.dbo.syslocks See Also sp_who, KILL sp_logdevice ──────────────────────────────────────────────────────────────────────────── Function Puts the syslogs system table, which contains the transaction log, on a separate database device. Syntax sp_logdevice dbname, database_device Examples create database products on default = 10, logs = 2 <execute> sp_logdevice products, logs <execute> Options dbname The name of the database whose syslogs table you want to put on a specific logical device. database_device The logical name of the database device you want to put the syslogs table on. This database device must be listed in sysdevices (run sp_helpdevice for a report on the existing database devices) and have been named in the CREATE DATABASE statement that the database dbname was defined in. Comments Whenever you create a database larger than about 4 megabytes, you must assign it to at least two separate logical database devices with the CREATE DATABASE statement and then use sp_logdevice to put the database's syslogs table on one of these database devices. This is essential to guarantee that SQL Server's automatic recovery mechanism works correctly. Putting the transaction log on a database device different from the rest of the database also improves performance. A smaller database can be created on a single database device, and the transaction log can be stored together with the rest of the database if you use only the DUMP DATABASE statement and never DUMP TRANSACTION for backups. The size of the database device required for the transaction log varies according to the amount of update activity and the frequency of transaction log dumps. As a rule of thumb, allocate to the log device 10 percent to 25 percent of the space you allocate to the database itself. It is best to start small, since space allocated to a transaction log device cannot be reclaimed. To increase the amount of allocated storage space on the transaction log database device, give the logical name of the log database device in the ON clause when you execute the ALTER DATABASE statement. Then execute sp_logdevice to cause the newly allocated space to be made available for the log. The database device on which you put syslogs is used only for the syslogs table. If you want to increase the amount of storage space allocated for the rest of the database, specify any database device other than the log database device when you execute the ALTER DATABASE statement. Use the DISK INIT statement to format a new physical database device to store databases and transaction logs on. If you run the DBCC CHECKALLOC statement after executing sp_logdevice, you will still see some pages for syslogs allocated on the database device. These pages are released for general database use only after they have been used by the transaction log and then dumped with DUMP TRANSACTION. After that, the transaction log will have been completely transferred to the database device named when you executed sp_logdevice. Only the owner of the database affected by sp_logdevice can execute it, but it can be executed in any database. See the SQL Server System Administrator's Guide for details. Messages No such database -- run sp_helpdb to list databases. No database with the supplied name exists. Run sp_helpdb to get a list of databases. No such device -- run sp_helpdevice to list SQL Server devices. The logical_name database device doesn't exist on SQL Server. Only the owner of the database may move the syslogs table. You aren't the owner of the dbname database. The specified device is not used by the database. The dbname database has no space allocated on the specified database device. Syslogs moved. The system procedure was successful, and the syslogs table is now located on the devname device. Permissions Execute permission defaults to the Database Owner of the named database. Tables Used master.dbo.sysdatabases, master.dbo.sysdevices, master.dbo.sysusages See Also sp_helpdevice, ALTER DATABASE, CREATE DATABASE, DISK INIT, DUMP DATABASE, DUMP TRANSACTION sp_monitor ──────────────────────────────────────────────────────────────────────────── Function Displays statistics about SQL Server. Syntax sp_monitor Examples sp_monitor This statement reports information about how busy SQL Server has been: last_run current_run seconds ------------------ -------------------- ----------- Sep 16 1987 4:52PM Sep 16 1987 4:53PM 55 (0 rows affected) cpu_busy io_busy idle ----------------- --------------- ------------------- 138(0)-0% 4(0)-0% 0(0)-0% (0 rows affected) ╓┌─┌──────────────────┌──────────────────┌──────────────────┌────────────────╖ ──────────────────────────────────────────────────────────────────────────── packets_received packets_sent packet_errors ----------------- ----------------- ----------------- 720(3) 806(4) 3(0) (0 rows affected) total_read total_write total_errors connections ----------------- ----------------- ----------------- ------------ ──────────────────────────────────────────────────────────────────────────── ----------------- ----------------- ----------------- ------------ 702(1) 23275(6) 0(0) 74(2) (0 rows affected) Options This system procedure has no options. Comments SQL Server keeps track of how much work it has done in a series of global variables. Executing sp_monitor displays the current values of these global variables and how much they have changed since the last time the procedure was run. For each column, the statistic is printed in the form number(number)-number% or number(number). The first number refers to the number of seconds (for cpu_busy, io_busy, and idle) or the total number (for the other variables) since SQL Server was restarted. The number in parentheses refers to the number of seconds or total number since the last time sp_monitor was run. The percentage is the percent of time since sp_monitor was last run. For example, if the report showed cpu_busy as 4250(215)-68%, this would mean that the CPU had been busy 4250 seconds since SQL Server was last started up, 215 seconds since sp_monitor was last run, and 68% of the total time since sp_monitor was last run. The following are the columns in the sp_monitor report and their meanings: last_run The time the sp_monitor system procedure was last run. current_run The time the sp_monitor system procedure is being run. seconds The number of elapsed seconds since the sp_monitor system procedure was run. cpu_busy The number of seconds that SQL Server's CPU was doing SQL Server work. io_busy The number of seconds that SQL Server has spent doing input and output operations. idle The number of seconds that SQL Server has been idle. pack_received The number of input packets read by SQL Server. pack_sent The number of output packets written by SQL Server. packet_errors The number of errors encountered by SQL Server while reading and writing packets. total_read The number of reads by SQL Server. total_write The number of writes by SQL Server. rdwr_errors The number of errors encountered by SQL Server while reading and writing. connections The number of logins or attempted logins to SQL Server. Permissions Execute permission defaults to the System Administrator. Tables Used master.dbo.spt_monitor See Also sp_who, Variables (Local and Global) sp_password ──────────────────────────────────────────────────────────────────────────── Function Adds or changes a password for an SQL Server login ID. Syntax sp_password old, new [, login_id] Examples A. sp_password null, ok, victoria The System Administrator has changed victoria's password from the default (NULL) to ok. Notice that NULL is not enclosed in quotation marks. B. sp_password ok, mediumhot The user victoria has changed her password from ok to mediumhot. Options old The user's old password, which is listed in master.dbo.syslogins. In the usual installation, only the System Administrator has SELECT permission on this field. The default password for all users is NULL. The System Administrator can give NULL for this parameter no matter what the user's old password was. However, if the System Administrator gives any value besides NULL, SQL Server checks it against master.dbo.syslogins and rejects it if it is wrong. new The user's new password. login_id An option that only the System Administrator can use to change any user's password. Comments Any user can change his or her password with sp_password. Only the System Administrator can use the login_id parameter and change someone else's password. The old parameter is checked against the existing password for the user login_id (unless the user executing the procedure is the System Administrator and the old parameter is NULL) and must match for the password to be changed. Passwords are listed in master.dbo.syslogins.password. As permissions are set up on the SQL Server distribution media, only the System Administrator has SELECT permission on this field. To give the NULL password in response to the Password: prompt, press ENTER. To give the NULL password with isql and the /P option, use isql /P. Messages Password changed. The password was successfully changed. The new password will be required the next time the user logs in to SQL Server. Only System Administrator can use loginame option -- password not changed. Any user can change his or her own password, but only the System Administrator can change another user's password. You aren't the System Administrator. No such login -- no password changed. The name supplied for the login_id parameter does not exist on SQL Server. Old (current) password incorrect for user -- password not changed. The password supplied as the old parameter is not the current password. Permissions Execute permission defaults to the public group. Each member of the public group has permission to change only his or her own password. The System Administrator can change any user's password. Tables Used master.dbo.syslogins See Also sp_addlogin, sp_adduser, sp_defaultdb sp_primarykey ──────────────────────────────────────────────────────────────────────────── Function Defines a primary key on a table or view. Syntax sp_primarykey tabname, col1 [, col2, col3, ..., col8] Examples A. sp_primarykey authors, au_id Example A makes the au_id field the primary key of authors. B. sp_primarykey employees, lastname, firstname Example B makes the combination of the fields lastname and firstname the primary key of the employees table. Options tabname The name of the table or view on which to define the primary key. col1 The name of the first column that makes up the primary key. The primary key can consist of between one and eight columns. Comments Executing sp_primarykey adds the key to the syskeys table. Keys are created to make explicit a logical relationship that is implicit in your database design. The information can be used by an application program. To display a report on the keys that have been defined, execute sp_helpkey. Only the owner of a table or view can define its primary key. A table or view can have only one primary key. Primary keys have been assigned within the master and model databases. Messages No table exists with the supplied name. The table or view supplied as the tabname parameter doesn't exist in the current database. Only the owner of the table may define a primary key. You aren't the owner of the table or view and therefore cannot define its primary key. Primary key already exists on table -- drop key first. A table or view can have only have one primary key; one already exists on the table or view supplied as the tabname parameter. The table has no such first column. The column name supplied as one of the column names isn't a column in tabname. New primary key added. The primary key has been added to the syskeys table. Permissions Execute permission defaults to the tabname owner. Tables Used syscolumns, syskeys, sysobjects See Also sp_commonkey, sp_dropkey, sp_foreignkey, sp_helpjoins, sp_helpkey, CREATE TRIGGER sp_rename ──────────────────────────────────────────────────────────────────────────── Function Changes the name of a user-created object in the current database. Syntax sp_rename objname, newname Examples A. sp_rename titles, books Example A renames the titles table to books. B. sp_rename "books.title", name Example B renames the title column in the books table to name. C. sp_rename tid, bookid Example C renames the user-defined datatype tid to bookid. Options objname The original name of the user object (table, view, column, stored procedure, trigger, default, or rule) or datatype. If the object to be renamed is a column in a table, objname must be in the form "table.column". You can rename an object in the current database only, and only if you own it. This rule holds for the Database Owner and System Administrator as well as for other users. newname The new name of the object or datatype. Names of objects and datatypes must follow the rules for identifiers. Comments Executing sp_rename changes the name of a user-created object or datatype. The names of most system datatypes and system objects cannot be changed. ──────────────────────────────────────────────────────────────────────────── WARNING Although it is possible to rename system procedures and the model and master databases, doing so will cause serious system problems. ──────────────────────────────────────────────────────────────────────────── You can change the name of an object or datatype in the current database only. A user can change the names only of those objects he or she owns. The Database Owner can change the name of any user's objects. ──────────────────────────────────────────────────────────────────────────── WARNING Procedures, triggers, and views that depend on an object whose name has been changed work fine until they are recompiled. However, recompilation takes place for many reasons and without notification to the user. When the procedure, trigger, or view is recompiled by SQL Server, it will no longer work. The user must change its text to reflect the new name of the object on which it depends. The safest course is to change the definitions of any dependent objects when you execute sp_rename. You can find dependent objects with sp_depends. ──────────────────────────────────────────────────────────────────────────── Messages Object must be in the current database. The name supplied for the objname parameter included a reference to a database. The object must be in the current database. Object name cannot be changed either because it does not exist in this database, or you don't own it, or it is a system name. The specified object is not an object or user-defined datatype, or you don't own the object. You do not own a table or column of that name in the current database. No column of the specified name exists in the specified table, or you don't own the table. Column name has been changed. The specified column name has been renamed to newname. Name of user-defined type name changed. The specified user-defined datatype has been renamed to newname. Newname already exists in sysobjects. The newname you specified is an existing name in the sysobjects table. Use a different newname. Object name has been changed. The specified object has been renamed to newname. Permissions Execute permission defaults to the object owners for their own objects. Execute permission defaults to the Database Owner for all objects. Tables Used syscolumns, sysobjects, systypes See Also sp_addtype, sp_depends, sp_renamedb, ALTER TABLE, CREATE DEFAULT, CREATE PROCEDURE, CREATE RULE, CREATE TABLE, CREATE TRIGGER, CREATE VIEW sp_renamedb ──────────────────────────────────────────────────────────────────────────── Function Changes the name of a database. Syntax sp_renamedb oldname, newname Examples sp_renamedb accounting, financial Renames the accounting database to financial. Options oldname The original name of the database. newname The new name of the database. Names of objects and datatypes must follow the rules for identifiers. Comments Executing sp_renamedb changes the name of a database. Only the System Administrator can change the name of a database. ──────────────────────────────────────────────────────────────────────────── WARNING Procedures, triggers, and views that depend on a database whose name has been changed work fine until they are recompiled. However, recompilation takes place for many reasons and without notification to the user. When the procedure, trigger, or view is recompiled by SQL Server, it will no longer work. The user must change its text to reflect the new name of the database on which it depends. The safest course is to change the definitions of any dependent objects when you execute sp_renamedb. ──────────────────────────────────────────────────────────────────────────── Messages Only the System Administrator can change the name of a database. Only the System Administrator can change database names, and you aren't the System Administrator. No specified database currently exists. The database you specified for the oldname parameter is not currently a database. A database with the new name already exists. The database you specified for the newname parameter is already a database. Database names must be unique, so you must use another name. Database name changed. The database name was successfully changed. Permissions Execute permission defaults to the object owners for their own objects. Execute permission defaults to the Database Owner for all objects. Tables Used master.dbo.sysdatabases See Also sp_changedbowner, sp_depends, sp_helpdb, sp_rename, CREATE DATABASE sp_spaceused ──────────────────────────────────────────────────────────────────────────── Function Displays the number of rows, the number of data pages, and the disk space used by an object or by each object in a database. Syntax sp_spaceused [objname] Examples A. sp_spaceused titles Example A reports on the amount of space allocated (reserved) for the titles table, the amount used for data, the amount used for index(es), and the available (unused) space: name rows reserved data index_size unused ------ ---- -------- ---- ---------- ------- titles 18 48 KB 6 KB 4 KB 38 KB B. sp_spaceused Example B prints a summary of space used in the database: database_name database_size ------------- -------------------- master 5 MB reserved data index_size unused --------- -------- ---------- ------- 2176 KB 1374 KB 72 KB 730 KB Options objname The name of the table or index under being examined. If the parameter is omitted, a summary of space used in the database is displayed. Comments Executing sp_spaceused computes the number of rows, the number of data pages, and the disk space used by the object or by each object in the database. The system procedure looks for an object in the current database. After you drop an index, sp_spaceused does not report accurate information. This is a known problem. Messages Object does not exist. The name supplied for the objname parameter is not an object in the current database. Permissions Execute permission defaults to the public group. Tables Used master..sysusages, sysindexes See Also sp_help, CREATE INDEX, CREATE TABLE, DROP INDEX, DROP TABLE sp_unbindefault ──────────────────────────────────────────────────────────────────────────── Function Unbinds a default value from a column or from a user-defined datatype. Syntax sp_unbindefault objname [, futureonly] Examples A. sp_unbindefault "employees.startdate" Example A unbinds the default from the startdate column of the employees table. B. sp_unbindefault ssn Example B unbinds the default from the user-defined datatype named ssn and all columns of that type. C. sp_unbindefault ssn, futureonly If you don't want to unbind defaults from existing ssn columns, use the string futureonly as an optional second parameter. Now the user datatype named ssn no longer has a default, but existing columns of type ssn are unaffected. Options objname The name of the table and column or datatype from which the default is to be unbound. If the parameter is not of the form "table.column", then objname is taken to be a usertype. When unbinding a default from a user datatype, any columns of that type that have the same default as the user datatype are also unbound. Columns of that type whose default has been changed are unaffected. futureonly Prevents existing columns of the specified user datatype from losing their defaults. Comments Executing sp_unbindefault removes a default from a column or from a user datatype in the current database. Columns of the user-defined datatype lose their current default unless their default has previously been changed or the value of the optional second parameter is futureonly. Messages Default unbound from table column. The table column supplied for the objname parameter no longer has any default. Default unbound from datatype. The user-defined datatype supplied for the objname parameter no longer has any default. You do not own a table with a column of that name. The table name supplied for the objname parameter either doesn't exist in the database or you don't own it. You can only bind or unbind defaults from columns in tables that you own. You do not own a user datatype of that name. The user-defined datatype supplied for the objname parameter either doesn't exist in the database or you don't own it. You can only bind or unbind defaults from datatypes that you own. Columns of the user datatype specified had their defaults unbound. Defaults on other columns of the user datatype specified were unbound unless their defaults had previously been changed. Column or user datatype must be in 'current' database. The objname parameter cannot include a database reference. Permissions Execute permission defaults to the object owner. Tables Used syscolumns, sysobjects, systypes See Also sp_bindefault, sp_bindrule, sp_unbindrule, CREATE DEFAULT, DROP DEFAULT sp_unbindrule ──────────────────────────────────────────────────────────────────────────── Function Unbinds a rule from a column or user-defined datatype. Syntax sp_unbindrule objname [, futureonly] Examples A. sp_unbindrule "employees.startdate" Example A unbinds the rule from the startdate column of the employees table. B. sp_unbindrule ssn Example B unbinds the rule from the user-defined datatype named ssn. If you don't want to unbind the rule from existing ssn columns, use the string futureonly as an optional second parameter. C. sp_unbindrule ssn, futureonly The user datatype ssn no longer has a rule, but no existing ssn columns are affected. Options objname The name of the table and column or user-defined datatype from which the rule is to be unbound. If the parameter is not of the form "table.column", then objname is taken to be a usertype. When unbinding a rule from a user datatype, any columns of that type that have the same rule as the user datatype are also unbound. Columns of that type whose rule has been changed are unaffected. futureonly Prevents existing columns of the specified user datatype from losing their rule. Comments Executing sp_unbindrule removes a rule from a column or user datatype in the current database. If unbinding a rule from a table column, the objname parameter must be of the form "table.column". The rule is unbound from all existing columns of the user-defined datatype unless their default has previously been changed or the value of the optional second parameter is futureonly. Messages Rule unbound from table column. The table column supplied for the objname parameter no longer has any rule. Rule unbound from datatype. The user-defined datatype supplied for the objname parameter no longer has any rule. You do not own a table with a column of that name. The table name supplied for the objname parameter either doesn't exist in the database or you don't own it. You can only bind or unbind rules on tables that you own. You do not own a user datatype of that name. The user-defined datatype supplied for the objname parameter either doesn't exist in the database or you don't own it. You can only bind or unbind rules from datatypes that you own. Columns of the user datatype specified had their rules unbound. Rules on other columns of the user datatype specified were unbound unless their rules had previously been changed. Column or user datatype must be in 'current' database. The objname parameter cannot include a database reference. Permissions Execute permission defaults to the object owner. Tables Used syscolumns, sysobjects, systypes See Also sp_bindefault, sp_bindrule, sp_unbindefault, CREATE RULE, DROP RULE sp_who ──────────────────────────────────────────────────────────────────────────── Function Reports information on current SQL Server users and processes. Syntax sp_who [login_id | "spid"] Examples A. sp_who Example A reports on the processes that are running on SQL Server: ╓┌───────┌──────────┌──────────┌──────────┌──────────┌───────┌──────────┌──── ───────────────────────────────────────────────────────────────────────────── spid status login_id hostname blk dbname cmd ----- --------- -------- -------- --- ------- ------ 1 sleeping sa plum 0 pubs AWAITI 2 sleeping sa 0 master NETWOR 3 sleeping sa 0 master CHECKP ───────────────────────────────────────────────────────────────────────────── 3 sleeping sa 0 master CHECKP 4 sleeping karenp plum 1 pubs SELECT 5 runnable karenp plum 0 pubs SELECT (5 rows affected) The spid column contains process identification numbers used in the TRANSACT-SQL KILL statement. The blk column contains the process IDs of the blocking process, if there is one. A blocking process (which can be infected or have an exclusive lock) is one that is holding resources that another process needs. In the previous example, process 4 (a SELECT on a table) is blocked by process 1 (a BEGIN TRANSACTION followed by an INSERT on the same table). B. sp_who victoria Example B reports on the processes the user victoria is running. C. sp_who "17" Example C reports on what SQL Server process number 17 is doing. Options login_id The user's login ID on SQL Server. If no name is specified, the procedure reports on all active users of SQL Server. spid The number of a specific process. You can supply a process number as a parameter if you enclose it in quotation marks (SQL Server is expecting a char type). Comments Executing sp_who reports information on a specified user, on a specified SQL Server process, or on all users currently running a process on SQL Server. Without parameters, the procedure reports which users are running what processes in all databases. The System Administrator can remove infected processes with the KILL statement. Messages No login exists with the supplied name. The name supplied for the login_id parameter does not exist on SQL Server. Permissions Execute permission defaults to the public group. Tables Used sysprocesses See Also sp_lock, KILL Chapter 3 Utility Programs ──────────────────────────────────────────────────────────────────────────── This chapter consists of reference pages, in alphabetical order, for the SQL Server utility programs. Each utility program and its function is listed below: ──────────────────────────────────────────────────────────────────────────── bcp Copies a database table to or from an operating system file in a user-specified format. bldmastr Builds the master database. console Dumps (or backs up) or loads (or restores) from a diskette drive. defncopy Copies definitions for specified views, rules, defaults, triggers, procedures, or reports from a database to an operating system file or from an operating system file to a database. isql Allows you to enter SQL statements and system procedures. sqlservr Starts SQL Server. ──────────────────────────────────────────────────────────────────────────── Also, the startsql utility program is called by DB-LIBRARY; users cannot call it directly. bcp ──────────────────────────────────────────────────────────────────────────── Function Copies a database table to or from an operating system file in a user-specified format. Syntax bcp /U login_id [[database_name.]owner.]table_name {in | out} datafile [/m maxerrors] [/f formatfile] [/e errfile] [/F firstrow] [/L lastrow] [/b batchsize] [/n] [/c] [/t field_term] [/r row_term] [/i inputfile] [/o outputfile] [/P password] [/S servername] [/v] Options /U login_id Allows the user to specify a login ID. database_name The database name, which is optional if the table being copied is in your default database. Otherwise, you must specify a database name. owner The owner's name, which is optional only if you own the table being copied. If no owner is specified and you do not own a table of that name, the program will not execute. table_name The name of the database table you want to copy. in, out The direction of the copy. The in option is a copy from a file into the database table, while the out option is a copy to a file from the database table. datafile The full pathname of an operating system file when copying a table to or from a hard disk file or a single diskette. The pathname can be from 1 to 255 characters in length. To copy a table to or from multiple diskettes, datafile is a drive specifier only (such as a:). /m maxerrors The maximum number of errors before the copy is aborted. Each row that can't be rebuilt by bcp is thrown out and counted as one error. The default used if this option is not included is 10. /f formatfile The full pathname of a file with stored responses from a previous use of bcp on the same table; creation of the format file is optional. Use this option when you have already created a format file that you want to use for a copy in or out. After you answer bcp's format questions, it will ask you if you want to save your answers in a format file. The default filename is bcp.fmt. The bcp program can refer to a format file when copying data so that you do not have to duplicate your previous format responses interactively. If this option is not used, bcp queries you for format information interactively. /e errfile The full pathname of an error file where bcp stores any rows that it was unable to transfer from the file to the database. Error messages from the bcp program go to the user's workstation. If this option is not used, no error file is created. /F firstrow The number of the first row to copy (the default is the first row). /L lastrow The number of the last row to copy (the default is the last row). /b batchsize The number of rows per batch of data copied (the default copies all the rows in one batch). /n Performs the copy operation using the data's native (database) datatypes as the default. This option does not prompt for each field; it uses the default values. /c Performs the copy operation with character type as the default. This option does not prompt for each field; it uses char as the default storage type, no prefixes, \t as the default field separator, and \n as the default row terminator. /t field_term Specifies the default field terminator. /r row_term Specifies the default row terminator. /i inputfile Allows the user to specify the name of a file that redirects input to bcp. /o outputfile Allows the user to specify the name of a file that receives output redirected from bcp. /P password Allows the user to specify a password. If the /P option is not used, bcp prompts for a password. If the /P option is used at the end of the command line without any password, bcp uses the default password (NULL). /S servername Allows the user to specify the name of the SQL Server to connect to. The servername is the name of the server computer on the network. This option is required when you are executing bcp from a remote computer on the network. /v Reports the current version of the bcp program. For more information and examples of how to use bcp see the SQL Server System Administrator's Guide. bldmastr ──────────────────────────────────────────────────────────────────────────── Function Builds the master database. Syntax bldmastr [ /d database_device ] [ /c cntrltype ] [ /s size ] [ /C ] [ /r ] [ /m ] Options /d database_device The master database device to be configured. /c cntrltype The controller number for the master database device. Together, cntrltype and disk specify the device. By default, a master database device always sets the cntrltype to 0. /s size The size of the master database device in 2K blocks. The bldmastr program verifies that the master database device has at least 2K blocks unless the /m or /r option is used. The default size is 10 megabytes (or 5120 2K blocks). /C Tells bldmastr to build a master database device that is not case sensitive. The default is case sensitive. See "Identifiers" in Chapter 1 for more information. /r Rewrites the configuration block that contains the system startup parameters with the default values without disturbing anything else contained on the disk. This option increments the version number without rebuilding the master database. /m Rewrites the master database only without changing the configuration block or reinitializing the master database device. If the /r and /m options are both used, then both the configuration block that contains the system start parameters (as derived from the system tables sysconfigures and syscurconfigs) and the master database are rewritten without initializing the rest of the disk. Comments The bldmastr program initializes the specified database device as an SQL Server master database device and builds the master and model databases on it. The bldmastr program also provides options for recovery from configuration errors and other disasters. The bldmastr program prompts for options that are not provided on the command line. If you don't supply any options, bldmastr prompts for each option as follows: master database name? master device controller number? master database size? master database starting offset? configuration only? (y or n) (same as /r) databases only? (y or n) (same as /m) case insensitive database? (y or n) In most cases, it is best to specify all bldmastr options on the command line rather than using bldmastr prompts. If you do not specify all options on the command line, be aware of the following limitations: ■ If you use the /d option only, the bldmastr program does not prompt you for other options, and it uses default values for /c, /s, and /C. ■ If you omit the /s option, the default size value is used, and you are not prompted for size. Files The bldmastr program contains all of the data needed to build the master and model databases. However, it does not include stored procedures, permissions, and so on. These are loaded during installation. console ──────────────────────────────────────────────────────────────────────────── Function Dumps or loads from a diskette drive. The console program is used by the person responsible for dumping and loading a database. ──────────────────────────────────────────────────────────────────────────── NOTE The console utility program is not available under Microsoft(R) Disk Operating System (MS-DOS(R)). ──────────────────────────────────────────────────────────────────────────── Syntax console [/S servername] The console program must be running before a DUMP DATABASE or LOAD DATABASE statement can proceed. See the "Comments" section. Options /S servername Allows the user to specify the name of the SQL Server to connect to. The servername is the name of the server computer on the network. This option is required when you are executing console from a remote computer on the network. Messages The following are the console's prompts when writing to a diskette (dumping a database or transaction log) and their explanations: ■ Prompt to mount diskette when writing: Mount first diskette: mount diskette on drive device_name. Type n to quit, press ENTER when mounted. device_name Physical name of the diskette dump device ■ Prompt to change diskette when writing (only when a diskette has been filled): Change diskette: mount diskette on drive device_name. Type n to quit, otherwise press ENTER when mounted. device_name Physical name of the diskette dump device ■ Prompt to remove the last diskette: Remove the diskette - press ENTER when done. The following are console's prompts when reading from diskette (loading a database or transaction log) and their explanations: ■ Prompt to mount diskette when reading: Mount first diskette: mount diskette_name volume_number on drive device_name. Type n if no diskettes are available, otherwise press ENTER when mounted. ■ Prompt to change diskette when reading: Change diskette: mount database_name volume_number on drive device_name. Type n if no diskettes are available, otherwise press ENTER when mounted. diskette_name Usually name of the dump, first 17 characters of database_name volume_number Diskette volume sequence number device_name Physical diskette dump device name, for example, a:\sqltable.dat ■ Prompt for operator override when reading: Diskette diskette_nameM volume_number mounted instead of diskette_name volume_number. Type n if you want to change diskette, otherwise press ENTER to use this one. The first diskette name and volume number are read from the diskette that was mounted. The second diskette name and volume number refer to the diskette that was expected to be mounted. Mounting a diskette out of sequence or with a different diskette name is usually an error, but an operator might have a good reason to override this error and use the diskettes in the order mounted. Mounting diskettes that were made on different days as part of the same load is usually an error, but an operator might have a good reason to do so. Comments The console program is a process that supports DUMP and LOAD for diskette drives. Relevant statements include DUMP DATABASE, DUMP TRANSACTION, LOAD DATABASE, and LOAD TRANSACTION. The console program prompts the operator for information relating to mounting and unmounting diskettes. The console program can be started each time SQL Server is started and left running in another screen group continuously. If the console program stops for any reason, the operator is prompted to restart it. Do not attempt to stop the console program with CONTROL+C or any other interrupt signal from the operating system. If SQL Server stops, console automatically stops. Since the console workstation (which can be any screen group) prints prompts and messages for the operator, you may want to run it in a window on a Microsoft Operating System/2 (MS(R) OS/2) workstation located close to the diskette unit. Most of the console program's messages are self-explanatory. To stop a dump or load to a diskette, type n in response to any prompt to mount a new diskette. If two diskettes (that is, two dumps/loads) are being used at once, only one console is needed. The program prints messages one at a time, waiting for a response before going to the next prompt. For this reason, the messages include the name of the physical drive to which they pertain. Make sure that the diskette you are using is formatted and not write-protected. The console program is not available under MS-DOS. ──────────────────────────────────────────────────────────────────────────── WARNING If the console utility program has been started and appears to be unusable, wait for up to one minute and try to use it again. (SQL Server may be in the process of fixing a broken console connection.) ──────────────────────────────────────────────────────────────────────────── defncopy ──────────────────────────────────────────────────────────────────────────── Function Copies definitions for specified views, rules, defaults, triggers, or procedures from a database to an operating system file or from an operating system file to a database. ──────────────────────────────────────────────────────────────────────────── NOTE The defncopy program cannot copy table definitions. ──────────────────────────────────────────────────────────────────────────── Syntax defncopy /U login_id [ /P password ] [ /S servername] { in filename dbname | out filename dbname [owner.]objectname [[owner.]objectname]...} Options /U login_id Allows the user to specify a login ID. /P password Allows the user to specify a password. If /P has no parameter, defncopy uses the default password (NULL); put /P by itself at the end of the command line. If /P is not used, the user is prompted for a password. /S servername Allows the user to specify the name of SQL Server to connect to. The servername is the name of the server computer on the network. This option is required if you are executing defncopy from a remote computer on the network. in, out Options that specify the direction of definition copy. filename The name of the operating system file destination or source for the definition copy. The copy out will overwrite an existing file. dbname The name of the database from or to which the definitions are being copied. objectname The name(s) of the database object(s) for defncopy to copy out. This parameter is not used when definitions are copied in. Comments The defncopy program is invoked directly from the operating system. It provides a noninteractive way of copying out definitions (CREATE statements) for views, rules, defaults, triggers, or procedures from a database to an operating system file. Alternatively, it will copy in all the definitions in a specified file. The infile or outfile name and the database name are required and must be unambiguous. For copying out, use filenames that reflect both the object's name and its owner. The defncopy program cannot be used to copy table definitions. You can copy out objects' definitions even if you do not have SELECT permission on these objects. (However, you do need SELECT permission on the sysobjects and syscomments tables for the relevant database.) You must have the appropriate CREATE permission for the type of object you are copying in. Objects copied in belong to the copier. A System Administrator copying in definitions on behalf of a user should be logged in as that user for that user to have proper access to the reconstructed database objects. A comment with more than 125 characters placed before a CREATE statement can cause defncopy to fail. isql ──────────────────────────────────────────────────────────────────────────── Function Allows you to enter SQL statements and system procedures. Syntax isql /U login_id [/e] [/p] [/i] [/n] [ /c cmdend] [ /h headers] [ /w columnwidth] [ /s colseparator] [ /t timeout] [ /m errorlevel] [ /H wksta_name] [ /P password] [/S servername] [/i inputfile] [/o outputfile] When using isql, you can use the following commands: go Terminates a command. reset Clears any statements you have entered. ed Calls the editor. !! command Executes an operating system command. quit, exit, or CONTROL+C Exits from isql. Options /U login_id Allows the user to specify a login ID. Login IDs are case sensitive. /e Echos input. /p Prints out performance statistics. /n Removes numbering and the prompt symbol (>) from input lines. /c cmdend Allows the user to reset the command terminator. By default, commands are terminated and sent to SQL Server by entering go on a line by itself. When you reset the command terminator, do not use SQL reserved words or characters that have special meaning to the operating system, whether preceded by a backslash or not. /h headers Allows the user to specify how many rows to print between column headings. The default is to print headings only once for each set of query results. /w columnwidth Allows the user to set the screen width for output. The default is 80 characters. When an output line has reached its maximum screen width, it is broken into multiple lines. /s colseparator Allows the user to reset the column separator character, which is blank by default. To use characters that have special meaning to the operating system (for example, |, ;, &, <, >), precede them with a backslash (\). /t timeout Allows the user to specify the number of seconds before a command times out. If no timeout is specified, a command runs indefinitely; however, the default timeout for logging in to isql is 60 seconds. /m errorlevel Allows the user to customize the display of error messages. The message number, state, and error level are displayed for errors of the specified severity level or higher. Nothing is displayed for errors of levels lower than the specified level. /H wksta_name Allows the user to specify a workstation name, changing the value in the dynamic system table sysprocesses, if he or she logs in from a different computer than usual. If no workstation name is specified, the current computer name is assumed. /P password Allows the user to specify a password. If the /P option is not used, isql prompts for a password. If the /P option is used at the end of the command line without any password, isql uses the default password (NULL). Passwords are case sensitive. /S servername Allows the user to specify the name of the SQL Server to connect to. The servername is the name of the server computer on the network. This option is required if you are executing isql from a remote computer on the network. /i inputfile Allows the user to specify the name of the file that contains a batch of SQL statements or stored procedures. The less-than symbol (<) can be used in place of /I. /o outputfile Allows the user to specify the name of the file that receives output from isql. The greater-than symbol (>) can be used in place of /o. Examples A. isql /p /s# B. isql /U joe Password: 1>select * 2>from authors 3>where city = "Oakland" 4>go Examples A and B execute the command. C. isql /U joe /P abracadabra 1>select * 2>from authors 3>where city = "Oakland" 4>ed Example C puts you in a text file where you can edit the query. When you save the file, you are returned to isql. The query is displayed. Type go to execute it. D. isql /U alma /P 1>select * 2>from authors 3>where city = "Oakland" 4>reset 1>quit % In example D, the reset command clears any statements you have entered. The quit command returns you to the operating system. Comments The isql program is started directly from the operating system. It accepts SQL commands and sends them to SQL Server. The results are formatted and printed on standard output. The isql program exits with quit or exit. Terminate a command by typing a line beginning with the command terminator. The default command terminator is go. You can follow the command terminator with an integer to specify how many times the command should be run. For example, to cause this command to be executed 100 times, type select x = 1 go 100 The results are printed once, at the end of execution. The user can call an editor on the current query buffer by entering ed as the first word on a line. Operating system commands can also be executed by starting a line with two exclamation points (!!) followed by the command. The editor is defined in the EDITOR environment variable. The default is "edlin". The existing query buffer can be cleared by typing reset on a line by itself. It causes any pending input to be discarded. The command terminator (go by default), reset, ed, !!, exit, quit, and CONTROL+C are recognized only if they appear at the very beginning of a line, immediately following the isql prompt. Anything entered on that line after these keywords is disregarded by isql. To use isql interactively, give the isql command (and any of the options) at your operating system prompt. You can read in a file containing a query for execution by isql by typing a command similar to the one shown below: isql /U alma /P /i stores.qry The file must include a command terminator. The results are displayed on the user's workstation. You can read in a file containing a query and direct the results to another file by typing a command similar to the one shown below: isql /U alma /P /i titles.qry /o titles.res Case is significant for isql options. When using isql interactively, you can read an operating system file into the command buffer with :r filename. Do not include a command terminator in the file; enter the terminator interactively once you have finished editing. You can include comments in a TRANSACT-SQL statement submitted to SQL Server by isql. Open a comment with "/*". Close it with "*/". Comments can be nested. For example: select au_lname, au_fname /*retrieve authors' last and first names*/ from authors, titles, titleauthor where authors.au_id = titleauthor.au_id and titles.title_id = titleauthor.title_id /*this is a three-way join that links authors **to the books they have written.*/ ──────────────────────────────────────────────────────────────────────────── NOTE Do not include a go command within a comment. A go command within a comment produces an error message. ──────────────────────────────────────────────────────────────────────────── sqlservr ──────────────────────────────────────────────────────────────────────────── Function Starts SQL Server. Syntax sqlservr /dmaster_phys_name [/eerror_log_phys_name] [/m] Options /dmaster_phys_name The full pathname of the master database device. /eerror_log_phys_name The full pathname of the error log file for SQL Server system-level error messages. /m Starts SQL Server in single user mode. Comments Note that the names of the master database device and the error log file cannot be preceded with a blank. This program should only be used to start SQL Server in single-user mode. Appendix A System Tables ──────────────────────────────────────────────────────────────────────────── System tables define the structure of the database. All system tables are found in the master database. Some system tables are found in all databases; they are automatically created when the CREATE DATABASE statement is executed. The following system tables exist in all databases: ──────────────────────────────────────────────────────────────────────────── sysalternates One row for each SQL Server user mapped to a database user syscolumns One row for each column in a table or view, and for each parameter in a stored procedure syscomments One or more rows for each view, rule, default, trigger, and stored procedure, giving an SQL definition statement sysdepends One row for each procedure, view, or table that is referenced by a procedure, view, or trigger sysindexes One row for each clustered index, nonclustered index, and table with no indexes syskeys One row for each foreign or primary key; set by user (not maintained by SQL Server) syslogs The transaction log sysobjects One row for each table, view, stored procedure, rule, trigger default, log, and (in tempdb only) temporary object sysprocedures One row for each view, rule, default, trigger, and stored procedure, giving an internal definition sysprotects User permissions information syssegments One row for each segment (named collection of disk pieces) systypes One row for each system-supplied and user-defined datatype sysusers One row for each user allowed in the database ──────────────────────────────────────────────────────────────────────────── The following system tables exist in the master database only: ──────────────────────────────────────────────────────────────────────────── sysconfigures One row for each user-settable configuration option syscurconfigs Information about configuration options currently being used by SQL Server sysdatabases One row for each database on SQL Server sysdevices One row for each disk dump device, disk for database device, and disk partition for databases syslocks Information about active locks syslogins One row for each valid SQL Server user account sysmessages One row for each system error or warning sysprocesses Information about server processes sysusages One row for each disk piece allocated to a database ──────────────────────────────────────────────────────────────────────────── In the pages that follow, each system table is described in more detail. Each table's columns are listed and described, and their datatypes are given. Finally, each table's indexes and the system procedures that reference it are listed. The word reserved in the column description means that the column is not currently used by SQL Server. Permissions for use of the system tables can be controlled by the Database Owner, just like permissions on any other tables. The SQL Server setup program sets up permissions so that all users can read the system tables, except for a few fields, like syslogins.passwd. All direct updates on system tables are disallowed by default, however, even for the Database Owner, because SQL Server has supplied system procedures to make any updates and additions to system tables that are normally needed. You can change the default and allow direct updates to the system tables if for some reason it becomes necessary to modify them in a way that has not been anticipated by SQL Server (and that therefore cannot be accomplished with a system procedure). This is accomplished by resetting the configuration option called allow updates with the sp_configure system procedure and the RECONFIGURE statement. For information on sp_configure, see the SQL Server Language Reference. ──────────────────────────────────────────────────────────────────────────── WARNING Some entries in some master database tables should not be altered by any user under any circumstances. For example, do not access syslogs in any way. Doing so may make it impossible for SQL Server to recover correctly in case of a system failure. In addition, an attempt to delete all rows from syslogs puts SQL Server into an infinite loop that eventually fills up the entire database. ──────────────────────────────────────────────────────────────────────────── sysalternates (all databases) Description Contains one row for each SQL Server user mapped (or with an alias) to a user of the current database. When a user tries to access a database, SQL Server looks for a valid uid entry in sysusers. If none is found, it looks in sysalternates.suid. If the user's suid is found there, he or she is treated as the database user whose suid is listed in sysalternates.altsuid. When SQL Server is installed, there are no entries in sysalternates. Column Datatype Description ──────────────────────────────────────────────────────────────────────────── suid smallint Server user ID of user being mapped altsuid smallint Server user ID of user to whom another user is mapped ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on suid Referenced by System Procedures sp_addalias, sp_adduser, sp_changedbowner, sp_dropalias, sp_dropuser, sp_helpuser syscolumns (all databases) Description Contains one row for every column in every table and view, and a row for each parameter in a stored procedure. ╓┌─────────┌─────────────┌───────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── id int ID of the table to which this column belongs or of Column Datatype Description ──────────────────────────────────────────────────────────────────────────── id int ID of the table to which this column belongs or of the stored procedure with which this parameter is associated number smallint Subprocedure number when the procedure is grouped (0 for nonprocedure entries) colid tinyint Column ID status tinyint Indicates the unique position for bit columns and whether null values are legal in this column type tinyint Physical storage type; copied from systypes length tinyint Physical length of data; copied from systypes or supplied by user offset smallint Offset into the row where this column appears; if negative, variable-length column Column Datatype Description ──────────────────────────────────────────────────────────────────────────── negative, variable-length column usertype smallint User type ID; copied from systypes cdefault int ID of the stored procedure that generates default value for this column domain int ID of the stored procedure that contains the rule for this column name sysname Column name printfmt varchar(255) reserved ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on id, number, and colid Referenced by System Procedures sp_bindefault, sp_bindrule, sp_commonkey, sp_droptype, sp_foreignkey, sp_help, sp_helpjoins, sp_helprotect, sp_primarykey, sp_rename, sp_unbindefault, sp_unbindrule syscomments (all databases) Description Contains entries for each view, rule, default, trigger, and stored procedure. The text field contains the original SQL definition statements. Since text is often longer than 255 characters, entries often span rows. Each object can occupy up to 255 rows. ╓┌─────────┌─────────────┌───────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── id int Object ID to which this text applies number smallint Grouped (0 for nonprocedure entries) colid tinyint Column ID if this entry is a column; 0 otherwise texttype smallint 0 System-supplied comment (for views, rules, defaults, triggers, and stored procedures) 1 User-supplied comment (users can add entries that describe an object or column) language smallint reserved text varchar(255) Actual text of SQL definition statement ──────────────────────────────────────────────────────────────────────────── Column Datatype Description ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on id, number, colid, and texttype Referenced by System Procedures sp_helptext sysconfigures (master database only) Description Contains one row for each user-settable configuration option. It contains the configuration options that were defined before the latest SQL Server startup plus any dynamic configuration options that were set since the latest SQL Server startup. The contents of sysconfigures is as follows: ╓┌───────┌───────────┌───────┌────────────────────────────┌──────────────────╖ ──────────────────────────────────────────────────────────────────────────── config value comment status 101 5 Maximum recovery interval 1 in minutes 102 0 Allow updates to system 1 tables 103 10 Number of user connections 0 allowed 104 1518 Size of available physical 0 ──────────────────────────────────────────────────────────────────────────── 104 1518 Size of available physical 0 memory in 2k pages 105 10 Number of open databases 0 allowed among all users 106 5000 Number of locks for all 0 users 107 500 Number of open database 0 objects 108 20 Percent of remaining memory 0 used for procedure cache 109 0 Default fill factor 0 percentage 110 100 Average time slice per 0 process in milliseconds ──────────────────────────────────────────────────────────────────────────── process in milliseconds 111 2 Default database size in 0 megabytes 112 0 Media retention period in 0 days 113 0 Recovery flags 0 114 0 Serial Number 0 (14 rows affected) Column Datatype Description ──────────────────────────────────────────────────────────────────────────── config smallint Configuration variable number value int User-modifiable value for the variable (being used by SQL Server only if RECONFIGURE has been executed) comment varchar(255) Explanation of the configuration option status varchar(2) 1 dynamic, the variable takes effect when RECONFIGURE is executed 0 The variable takes effect when SQL Server is restarted ──────────────────────────────────────────────────────────────────────────── ──────────────────────────────────────────────────────────────────────────── NOTE The values shown above are created when you install SQL Server with the setup program. If you install SQL Server with the bldmastr utility program, the values column will be different. bldmastr creates a 0 in the values column to indicate that the default value is in effect. ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on config Referenced by System Procedures sp_configure syscurconfigs (master database only) Description Is built dynamically when queried. It contains an entry for each of the configuration options, as does sysconfigures, but with the current values. In addition, it contains four entries that describe the configuration structure. The contents of syscurconfigs is as follows: ╓┌───────┌───────────┌───────┌────────────────────────────┌──────────────────╖ ──────────────────────────────────────────────────────────────────────────── config value comment status 1 1 Major revision number of 0 config data 2 0 Minor revision number of 0 config data 3 1 Reconfigure revision number 0 of config data 4 2 Configuration boot source 0 101 5 Maximum recovery interval ──────────────────────────────────────────────────────────────────────────── 101 5 Maximum recovery interval in minutes 1 102 0 Allow updates to system 1 tables 103 10 Number of user connections 0 allowed 104 1518 Size of available physical 0 memory in 2k pages 105 10 Number of open databases 0 allowed among all users 106 5000 Number of locks for all 0 users 107 500 Number of open database 0 objects ──────────────────────────────────────────────────────────────────────────── objects 108 20 Percent of remaining memory 0 used for procedure cache 109 0 Default fill factor 0 percentage 110 100 Average time slice per 0 process in milliseconds 111 2 Default database size in 0 megabytes 112 0 Media retention period in 0 days 113 0 Recovery flags 0 114 666666 Serial Number 0 ──────────────────────────────────────────────────────────────────────────── 114 666666 Serial Number 0 (18 rows affected) Referenced by System Procedures sp_configure sysdatabases (master database only) Description Contains one row for each database on SQL Server. When SQL Server is installed, sysdatabases contains entries for the master database, the model database, and the tempdb temporary database. ╓┌────────┌─────────┌────────────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── name sysname Name of the database dbid smallint Database ID suid smallint Server user ID of database creator mode smallint Used internally for locking a database while it is being created status smallint Control bits, some of which can be set by the user with sp_dboption (read only, dbo only, single user, and so on): 0x40 select into/bulkcopy; set with sp_dboption 0x08 truncate log on chkpt; set with sp_dboption 0x10 no chkpt on recovery; set with sp_dboption Column Datatype Description ──────────────────────────────────────────────────────────────────────────── 0x10 no chkpt on recovery; set with sp_dboption 0x20 Crashed while database was being loaded; instructs recovery not to proceed 0x100 Database is suspect; cannot be opened or used in its present state and can be dropped only with DBCC DBREPAIR 0x400 Read only; set with sp_dboption 0x800 dbo use only; set with sp_dboption 0x1000 Single user; set with sp_dboption 0x4000 dbname has changed version smallint Version of SQL Server code under which database was created crdate datetime Creation date ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on name Unique nonclustered index on dbid Referenced by System Procedures sp_addlogin, sp_changedbowner, sp_dboption, sp_defaultdb, sp_depends, sp_helpdb, sp_logdevice, sp_renamedb sysdepends (all databases) Description Contains one row for each procedure, view, or table that is referenced by a procedure, view, or trigger. ╓┌──────────┌─────────┌──────────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── id int Object ID number smallint Procedure number depid int Dependent object ID depnumber smallint Dependent procedure number depdbid Reserved for future use depsiteid Reserved for future use status smallint Internal status information selall bit On if object is used in SELECT * statement resultobj bit On if object is being updated readobj bit On if object is being read ──────────────────────────────────────────────────────────────────────────── Indexes Clustered index on id, number, depid, depnumber, dpedbid, and depsiteid Referenced by System Procedures sp_depends sysdevices (master database only) Description Contains one row for each disk dump device, diskette dump device, and database device. When SQL Server is installed, there are four entries in sysdevices: one for the master database device (for databases), one for a disk dump device, and two for diskette dump devices. ╓┌──────────┌─────────────┌──────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── low int First virtual page number on database device (not used for dump devices) Column Datatype Description ──────────────────────────────────────────────────────────────────────────── used for dump devices) high int Last virtual page number on database device (not used for dump devices) status smallint Indicates whether it is to be used as a default: 2 Physical database device for database storage 3 Physical database device for default database storage 16 Dump device (either disk or diskette) 24 Dump device (either disk or diskette); skip ANSI labels cntrltype smallint Controller type (0 if database device, 2 if hard disk dump device, 3-4 if diskette dump device) name sysname Logical name of dump device or database device phyname varchar(127) Name of physical database device or dump device Column Datatype Description ──────────────────────────────────────────────────────────────────────────── phyname varchar(127) Name of physical database device or dump device ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on name Referenced by System Procedures sp_addumpdevice, sp_diskdefault, sp_dropdumpdevice, sp_helpdevice, sp_logdevice sysindexes (all databases) Description Contains one row for each clustered index, one row for each nonclustered index, and one row for each table that has no clustered index. The columns dpages, reserved, used, and rows are maintained only if the row describes a table or clustered index. ╓┌─────────────┌───────────────┌─────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── name sysname Index or table name id int ID of table or ID of table to which index belongs indid smallint 0 if table, 1 if clustered index, >1 if nonclustered Column Datatype Description ──────────────────────────────────────────────────────────────────────────── dpages int Number of data pages used (if entry is a table or clustered index) reserved int Number of pages allocated (if entry is a table or clustered index) used int Number of data and index pages (if entry is a clustered index) rows int Number of rows (if entry is a table or clustered index) first int Pointer to first data or leaf page root int Pointer to root page if entry is an index; pointer to last page if entry is a table distribution int Pointer to distribution page (if entry is an Column Datatype Description ──────────────────────────────────────────────────────────────────────────── distribution int Pointer to distribution page (if entry is an index) usagecnt int reserved segment smallint Number of segment in which this object resides status smallint Internal system status information: 0x1 Abort command if attempt to insert duplicate key 0x2 Unique index 0x4 Abort command if attempt to insert duplicate row 0x10 Clustered index 0x40 Index allows duplicate rows rowpage smallint Maximum number of rows per page Column Datatype Description ──────────────────────────────────────────────────────────────────────────── minlen smallint Minimum size of a row maxlen smallint Maximum size of a row maxirow smallint Maximum size of a nonleaf index row keycnt smallint Number of keys keys1 varbinary(255) Description of key columns (if entry is an index) keys2 varbinary(255) Description of key columns (if entry is an index) ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on id and indid Referenced by System Procedures sp_helpindex, sp_spaceused syskeys (all databases) Description Contains one row for each foreign or primary key. ╓┌────────┌─────────┌────────────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── Column Datatype Description ──────────────────────────────────────────────────────────────────────────── id int int null type smallint Record type depid int null Dependent object ID keycnt int Number of non-null keys size int null reserved key1 int null Column ID key2 int null Column ID key3 int null Column ID key4 int null Column ID key5 int null Column ID key6 int null Column ID key7 int null Column ID key8 int null Column ID depkey1 int null Column ID depkey2 int null Column ID depkey3 int null Column ID depkey4 int null Column ID depkey5 int null Column ID depkey6 int null Column ID Column Datatype Description ──────────────────────────────────────────────────────────────────────────── depkey6 int null Column ID depkey7 int null Column ID depkey8 int null Column ID ──────────────────────────────────────────────────────────────────────────── Indexes Clustered index on id Referenced by System Procedures sp_commonkey, sp_dropkey, sp_foreignkey, sp_helpkey, sp_helpjoins, sp_primarykey syslocks (master database only) Description Contains information about active locks. The syslocks system table is not a standard table. Rather, it is built dynamically when queried by a user. No updates to syslocks are allowed. ╓┌───────┌─────────┌─────────────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── id int Table ID dbid smallint Database ID page int Page number type smallint Type of lock: 1 Exclusive table lock 2 Shared table lock 3 Exclusive intent lock (will do page locking on Column Datatype Description ──────────────────────────────────────────────────────────────────────────── 3 Exclusive intent lock (will do page locking on indicated pages) 4 Shared intent lock 5 Exclusive page lock 6 Shared page lock 7 Update page lock (changes to exclusive if page is actually modified) 8 Exclusive extent lock 9 Shared extent lock 0x100 Lock is blocking another process 0x200 Demand lock spid smallint ID of process that holds the lock ──────────────────────────────────────────────────────────────────────────── Indexes None Referenced by System Procedures sp_lock syslogins (master database only) Description Contains one row for each valid SQL Server user account. When SQL Server is installed, syslogins contains one entry, in which the name is sa, the suid is 1, and the password is null. ╓┌────────────┌─────────────┌────────────────────────────────────────────────╖ Column Datatype Description Column Datatype Description ──────────────────────────────────────────────────────────────────────────── suid smallint Server user ID status smallint reserved accdate datetime reserved totcpu int reserved totio int reserved spacelimit int reserved timelimit int reserved resultlimit int reserved dbname sysname Name of database to put user into when connection is established Column Datatype Description ──────────────────────────────────────────────────────────────────────────── name sysname Login ID of user password sysname null Password of user (may be null) ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on suid Unique nonclustered index on name Referenced by System Procedures sp_addalias, sp_addlogin, sp_adduser, sp_changedbowner, sp_defaultdb, sp_defdb, sp_droplogin, sp_helpdb, sp_helpuser, sp_password syslogs (all databases) Description Contains the transaction log. It is used by SQL Server for recovery and roll-forward; it is not used by system users. You cannot delete from, insert into, or update syslogs. Every data modification operation is logged, so before you can change syslogs, the change must be logged. This means that a change operation on syslogs adds a row to syslogs, which then must be logged, adding another row to syslogs, and so on, producing an infinite loop. The loop continues until the database becomes full. Column Datatype Description ──────────────────────────────────────────────────────────────────────────── xactid binary(6) Transaction ID op tinyint Update operation number ──────────────────────────────────────────────────────────────────────────── Indexes None sysmessages (master database only) Description Contains one row for each system error or warning that can be returned by SQL Server. SQL Server displays the error description on the user's screen. Column Datatype Description ──────────────────────────────────────────────────────────────────────────── error int Unique error number severity smallint Severity level of error dlevel smallint Reserved for number of descriptive level of this message: terse, short, or long description varchar(255) Explanation of error with place holders for parameters ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on error and dlevel sysobjects (all databases) Description Contains one row for each table, view, stored procedure, log, rule, default, trigger, and (in tempdb only) temporary object. ╓┌─────────┌─────────┌───────────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── name sysname Object name id int Object ID uid smallint User ID of owner object type char One of the following object types: S System table U User table V View L Log P Stored procedure R Rule D Default Column Datatype Description ──────────────────────────────────────────────────────────────────────────── D Default TR Trigger userstat smallint Application-dependent type information sysstat smallint Internal status information indexdel smallint Index delete count (incremented if an index is deleted) schema smallint Count of changes in schema of a given object (incremented if a rule or default is added) refdate datetime reserved crdate datetime Date object was created expdate datetime reserved deltrig int Stored procedure ID of a delete trigger Column Datatype Description ──────────────────────────────────────────────────────────────────────────── deltrig int Stored procedure ID of a delete trigger instrig int Stored procedure ID of an insert trigger updtrig int Stored procedure ID of an update trigger seltrig int reserved category int reserved cache smallint reserved ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on id Unique nonclustered index on name and uid Referenced by System Procedures sp_bindefault, sp_bindrule, sp_commonkey, sp_depends, sp_dropgroup, sp_dropkey, sp_droptype, sp_dropuser, sp_foreignkey, sp_help, sp_helpjoins, sp_helprotect, sp_primarykey, sp_rename, sp_unbindefault, sp_unbindrule sysprocedures (all databases) Description Contains entries for each view, default, rule, trigger, and stored procedure. The plan or sequence tree for each object is stored in binary form. If the sequence tree doesn't fit in one entry, it is broken into more than one row. The sequence field identifies the sub-rows. ╓┌─────────┌─────────┌───────────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── type smallint Object type: 0x1 Entry describes a plan (reserved) 0x2 Entry describes a tree id int Object ID sequence smallint Sequence number if more than one row is used to describe this object status smallint Internal system status number smallint Subprocedure number when the procedure is grouped (0 for nonprocedure entries) ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on id, number, type, and sequence sysprocesses (master database only) Description Contains information about SQL Server processes. The sysprocesses system table is not a standard table. Rather, it is built dynamically when queried by a user. No updates to sysprocesses are allowed. Use the KILL statement to kill a process. ╓┌─────────────┌─────────┌───────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── spid smallint Process ID kpid smallint Kernel process ID Column Datatype Description ──────────────────────────────────────────────────────────────────────────── kpid smallint Kernel process ID status char(10) Process ID status (for example, runnable, sleeping, and so on) suid smallint Server user ID of user who executed command hostname char(10) Name of workstation program_name char(16) Name of application program hostprocess char(8) Workstation process ID number cmd char(16) Command currently being executed cpu int Cumulative CPU time for process physical_io int Number of disk reads and writes for current command Column Datatype Description ──────────────────────────────────────────────────────────────────────────── command memusage int Amount of memory allocated to process blocked smallint Process ID of blocking process, if any dbid smallint Database ID uid smallint ID of user who executed command gid smallint Group ID of user who executed command ──────────────────────────────────────────────────────────────────────────── Indexes None Referenced by System Procedures sp_who sysprotects (all databases) Description Contains user permissions information─entries for each GRANT and REVOKE statement that has been executed. ╓┌────────────┌──────────────┌───────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── id int ID of object to which this permission applies Column Datatype Description ──────────────────────────────────────────────────────────────────────────── uid smallint ID of user or group to which this permission applies action tinyint One of the following permissions: SELECT = 193 INSERT = 195 DELETE = 196 UPDATE = 197 EXECUTE = 224 CREATE DATABASE = 203 CREATE DEFAULT = 233 CREATE PROCEDURE = 222 CREATE RULE = 236 CREATE TABLE = 198 CREATE VIEW = 207 DUMP DATABASE = 228 DUMP TRANSACTION = 235 Column Datatype Description ──────────────────────────────────────────────────────────────────────────── protecttype tinyint Either 205 (grant) or 206 (revoke) columns varbinary(32) Bit map of columns to which this SELECT or UPDATE permission applies. Bit 0 indicates all columns; bit 1 means permission applies to that column; NULL means no information. ──────────────────────────────────────────────────────────────────────────── Indexes Clustered index on id, uid, and action Referenced by System Procedures sp_helprotect syssegments (all databases) Description Contains one row for each segment (named collection of disk pieces). When SQL Server is installed, this table contains two entries: segment 0 (system) for system tables and segment 1 (default) for other objects. All entries in sysusages contain both these segments in their maps. Column Datatype Description ──────────────────────────────────────────────────────────────────────────── segment smallint Segment number name sysname Segment name, used internally status int null Indicates which is the default segment ──────────────────────────────────────────────────────────────────────────── Indexes None systypes (all databases) Description Contains one row for each system-supplied and each user-defined datatype. Domains (defined by rules) and defaults are given if they exist. The rows that describe system-supplied datatypes cannot be altered. The system-supplied datatypes and their ID numbers (the contents of the name and type fields, respectively) are as follows: ╓┌───────────────────────────────┌───────────────────────────────────────────╖ System Datatype ID Number ──────────────────────────────────────────────────────────────────────────── binary 45 System Datatype ID Number ──────────────────────────────────────────────────────────────────────────── binary 45 bit 50 char 4 datetime 61 datetimn 111 float 62 floatn 109 image 46 int 56 intn 38 money 60 moneyn 110 smallint 52 sysname 39 text 35 timestamp 37 tinyint 48 varbinary 37 varchar 39 System Datatype ID Number ──────────────────────────────────────────────────────────────────────────── varchar 39 ──────────────────────────────────────────────────────────────────────────── The following table shows the datatype and description for each column in the systypes table. ╓┌───────────┌─────────────┌─────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── uid smallint User ID of datatype creator usertype smallint User type ID variable bit 1 if datatype is variable length; 0 otherwise allownulls bit Indicates whether NULLs are allowed for this datatype Column Datatype Description ──────────────────────────────────────────────────────────────────────────── type tinyint Physical storage datatype length tinyint Physical length of datatype tdefault int ID of stored procedure that generates default for this datatype domain int ID of stored procedure that contains integrity checks for this datatype name sysname Datatype name printfmt varchar(255) reserved ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on name Unique nonclustered index on usertype Referenced by System Procedures sp_addtype, sp_bindefault, sp_bindrule, sp_droptype, sp_dropuser, sp_help, sp_rename, sp_unbindefault, sp_unbindrule sysusages (master database only) Description Contains one row for each disk allocation piece assigned to a database. Each database contains a specified number of database (logical) page numbers. Each disk piece includes segments 0 and 1. When the CREATE DATABASE statement is executed, SQL Server scans the disk(s) to find available disk allocation pieces. One or more contiguous disk allocation pieces is assigned to the database, and the mapping is recorded in sysusages. ╓┌───────┌─────────┌─────────────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── dbid smallint Database ID segmap int Bit map of possible segment assignments lstart int First database (logical) page number size int Number of contiguous database (logical) pages vstart int Starting virtual page number ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on dbid and lstart Unique nonclustered index on vstart Referenced by System Procedures sp_logdevice sysusers (all databases) Description Contains one row for each user allowed in the database and one row for each group. On the SQL Server distribution diskette, master..sysusers contains three entries: dbo, whose suid is 1 and uid is 1; guest, whose suid is -1 and uid is 2; and public, whose suid is -2 and uid is 0. The sysusers table in the model database (and thus in all user databases) initially contains two entries: dbo and public. The user guest provides a mechanism for giving access to the database to users not explicitly listed in sysusers, with a restricted set of permissions. The guest entry in master means that any user with an account on SQL Server (that is, with an entry in syslogins) can access master. The user public refers to all users. The keyword PUBLIC is used with the GRANT and REVOKE statements to signify that the permission is being given to or taken away from all users. ╓┌────────┌─────────────┌────────────────────────────────────────────────────╖ Column Datatype Description ──────────────────────────────────────────────────────────────────────────── uid smallint Server user ID, copied from syslogins. suid 1 is the System Administrator; -1 is a guest account. uid smallint User ID, unique in this database, used for granting and revoking permissions. uid 1 is the dbo. gid smallint Group ID to which this user belongs. If uid = gid, this entry defines a group, and suid for this entry is 0. Column Datatype Description ──────────────────────────────────────────────────────────────────────────── is 0. name sysname Username or groupname, unique in this database. environ varchar(255) reserved ──────────────────────────────────────────────────────────────────────────── Indexes Unique clustered index on suid Unique nonclustered index on name Unique nonclustered index on uid Referenced by System Procedures sp_addalias, sp_adduser, sp_addgroup, sp_changedbowner, sp_changegroup, sp_dropgroup, sp_droplogin, sp_droptype, sp_dropuser, sp_helpgroup, sp_helprotect, sp_helpuser Figure A.1: System Tables Relationship Chart - (This figure may be found in the printed book.) Appendix B The pubs Sample Database ──────────────────────────────────────────────────────────────────────────── This appendix describes the sample database, pubs. The pubs database has eight tables: ■ publishers ■ titleauthor ■ sales ■ stores ■ discounts ■ roysched ■ authors ■ titles Each database table is described by two figures. The first figure documents the structure of the table; for each table field, the figure lists its datatype, its NULL/NOT NULL status, and any defaults, rules, triggers, and indexes. The second figure lists the table contents. Table publishers ─ Structure pub_id pub_name city state ──────────────────────────────────────────────────────────────────────────── Datatype char(4) varchar(40) varchar(20) char(2) Null not null null null null Rule pub_idrule(1) ─ ─ ─ Index clust, uniq ─ ─ ─ ──────────────────────────────────────────────────────────────────────────── (1) The pub_idrule states that the data must be 1389, 0736, 0877, 1622, or 1756, or must match the pattern 99[0-9][0-9]. pub_id pub_name city state ──────────────────────────────────────────────────────────────────────────── 1389 Algodata Berkeley CA Infosystem s 0736 New Age Boston MA Books 0877 Binnet & Washington DC Hardley Table titleauthor ─ Structure au_id title_id au_ord royaltyper ──────────────────────────────────────────────────────────────────────────── Datatype id tid smallint int Null not null not null null null Index nonclust nonclust ─ ─ uniq, nonclust, composite ──────────────────────────────────────────────────────────────────────────── Table titleauthor ─ Contents ╓┌────────────┌─────────┌───────┌────────────────────────────────────────────╖ au_id title_id au_ord royaltyper ──────────────────────────────────────────────────────────────────────────── 409-56-7008 BU1032 1 60 213-46-8915 BU1032 2 40 238-95-7766 PC1035 1 100 213-46-8915 BU2075 1 100 998-72-3567 PS2091 1 50 899-46-2035 PS2091 2 50 998-72-3567 PS2106 1 100 722-51-5454 MC3021 1 75 899-46-2035 MC3021 2 25 807-91-6654 TC3218 1 100 486-29-1786 PS7777 1 100 486-29-1786 PC9999 1 100 712-45-1867 MC2222 1 100 172-32-1176 PS3333 1 100 274-80-9391 BU7832 1 100 427-17-2319 PC8888 1 50 au_id title_id au_ord royaltyper ──────────────────────────────────────────────────────────────────────────── 427-17-2319 PC8888 1 50 846-92-7186 PC8888 2 50 756-30-7391 PS1372 1 75 724-80-9391 PS1372 2 25 724-80-9391 BU1111 1 60 267-41-2394 BU1111 2 40 672-71-3249 TC7777 1 40 267-41-2394 TC7777 2 30 472-27-2349 TC7777 3 30 648-92-1872 TC4203 1 100 ──────────────────────────────────────────────────────────────────────────── Table sales ─ Structure stor_id ord_num date qty payterms title_id ──────────────────────────────────────────────────────────────────────────── Datatype char(4) varchar(20) datetime smallint varchar(12) tid Null not null not null not null not null not null not null Index ─ ─ ─ ─ ─ nonclust ──────────────────────────────────────────────────────────────────────────── Table sales ─ Contents ╓┌────────┌─────────┌─────────┌────┌───────────┌─────────────────────────────╖ stor_id ord_num date qty payterms title_id ──────────────────────────────────────────────────────────────────────────── 7066 QA7442.3 09/13/85 75 On invoice PS2091 7067 D4482 09/14/85 10 Net 60 PS2091 7131 N914008 09/14/85 20 Net 30 PS2091 7131 N914014 09/14/85 25 Net 30 MC3021 8042 423LL922 09/14/85 15 On invoice MC3021 8042 423LL930 09/14/85 10 On invoice BU1032 6380 722a 09/13/85 03 Net 60 PS2091 6380 6871 09/14/85 05 Net 60 BU1032 8042 P723 03/11/88 25 Net 30 BU1111 7896 X999 02/21/88 35 On invoice BU2075 7896 QQ2299 10/28/87 15 Net 60 BU7832 7896 TQ456 12/12/87 10 Net 60 MC2222 8042 QA879.1 05/22/87 30 Net 30 PC1035 stor_id ord_num date qty payterms title_id ──────────────────────────────────────────────────────────────────────────── 8042 QA879.1 05/22/87 30 Net 30 PC1035 7066 A2976 05/24/87 50 Net 30 PC8888 7131 P3087a 05/29/87 20 Net 60 PS1372 7131 P3087a 05/29/87 25 Net 60 PS2106 7131 P3087a 05/29/87 15 Net 60 PS3333 7131 P3087a 05/29/87 25 Net 60 PS7777 7067 P2121 05/15/87 40 Net 30 TC3218 7067 P2121 05/15/87 20 Net 30 TC4203 7067 P2121 05/15/87 20 Net 30 TC7777 ──────────────────────────────────────────────────────────────────────────── Table stores ─ Structure stor_id stor_name stor_addres city state zip s ───────────────────────────────────────────────────────────────────────────── Datatype char(4) varchar(40) varchar(40) varchar(20) char(2) char(5) Null not null null null null null null ───────────────────────────────────────────────────────────────────────────── Table stores ─ Contents ╓┌────────┌────────────────────┌────────────────────┌──────────┌──────┌──────╖ stor_id stor_name stor_address city state zip ──────────────────────────────────────────────────────────────────────────── 7066 Barnum's 567 Pasadena Ave. Tustin CA 92789 7067 News & Brews 577 First St. Los Gatos CA 96745 7131 Doc-U-Mat: Quality 24-A Avrogado Way Remulade WA 98014 Laundry and Books 8042 Bookbeat 679 Carson St. Portland OR 89076 6380 Eric the Read Books 788 Catamaugus Ave. Seattle WA 98056 stor_id stor_name stor_address city state zip ──────────────────────────────────────────────────────────────────────────── 7896 Fricative Bookshop 89 Madison St. Fremont CA 90019 ──────────────────────────────────────────────────────────────────────────── B.1.1 discounts ─ Structure - discounttype stor_id lowqty highqty discount ──────────────────────────────────────────────────────────────────────────── Datatype varchar(40) char(4) smallint smallint float Null not null null null null not null discounts ─ Contents discounttype stor_id lowqty highqty discount ──────────────────────────────────────────────────────────────────────────── Initial Customer ─ ─ ─ 10.5 Volume Discount ─ 100 1000 6.7 Customer Discount 8042 ─ ─ 5.0 roysched ─ Structure - title_id lorange hirange royalty ──────────────────────────────────────────────────────────────────────────── Datatype tid int int int Null not null null null null Index nonclust ─ ─ ─ Table roysched ─ Contents ╓┌─────────┌────────┌─┌────────┌─┌───────────────────────────────────────────╖ title_id lorange hirange royalty ──────────────────────────────────────────────────────────────────────────── BU1032 0 5000 10 BU1032 5001 50000 12 PC1035 0 2000 10 PC1035 2001 3000 12 PC1035 3001 4000 14 PC1035 4001 10000 16 PC1035 10001 50000 18 BU2075 0 1000 10 BU2075 1001 3000 12 BU2075 3001 5000 14 BU2075 5001 7000 16 BU2075 7001 10000 18 BU2075 10001 12000 20 BU2075 12001 14000 22 BU2075 14001 50000 24 title_id lorange hirange royalty ──────────────────────────────────────────────────────────────────────────── BU2075 14001 50000 24 PS2091 0 1000 10 PS2091 1001 5000 12 PS2091 5001 10000 14 PS2091 10001 50000 16 PS2106 0 2000 10 PS2106 2001 5000 12 PS2106 5001 10000 14 PS2106 10001 50000 16 MC3021 0 1000 10 MC3021 1001 2000 12 MC3021 2001 4000 14 MC3021 4001 6000 16 MC3021 6001 8000 18 MC3021 8001 10000 20 MC3021 10001 12000 22 MC3021 12001 50000 24 TC3218 0 2000 10 TC3218 2001 4000 12 title_id lorange hirange royalty ──────────────────────────────────────────────────────────────────────────── TC3218 2001 4000 12 TC3218 4001 6000 14 TC3218 6001 8000 16 TC3218 8001 10000 18 TC3218 10001 12000 20 TC3218 12001 14000 22 TC3218 14001 50000 24 TC3218 0 2000 10 TC3218 2001 4000 12 TC3218 4001 6000 14 TC3218 6001 8000 16 TC3218 8001 10000 18 TC3218 10001 12000 20 TC3218 12001 14000 22 TC3218 14001 50000 24 PC8888 0 5000 10 PC8888 5001 10000 12 PC8888 10001 15000 14 PC8888 15001 50000 16 title_id lorange hirange royalty ──────────────────────────────────────────────────────────────────────────── PC8888 15001 50000 16 PS7777 0 5000 10 PS7777 5001 50000 12 PS3333 0 5000 10 PS3333 5001 10000 12 PS3333 10001 15000 14 PS3333 15001 50000 16 BU1111 0 4000 10 BU1111 4001 8000 12 BU1111 8001 10000 14 BU1111 12001 16000 16 BU1111 16001 20000 18 BU1111 20001 24000 20 BU1111 24001 28000 22 BU1111 28001 50000 24 MC2222 0 2000 10 MC2222 2001 4000 12 MC2222 4001 8000 14 MC2222 8001 12000 16 title_id lorange hirange royalty ──────────────────────────────────────────────────────────────────────────── MC2222 8001 12000 16 MC2222 8001 12000 16 MC2222 12001 20000 18 MC2222 20001 50000 20 TC7777 0 5000 10 TC7777 5001 15000 02 TC7777 15001 50000 14 TC4203 0 2000 10 TC4203 2001 8000 12 TC4203 8001 16000 14 TC4203 16001 24000 16 TC4203 24001 32000 18 TC4203 32001 40000 20 TC4203 40001 50000 22 BU7832 0 5000 10 BU7832 5001 10000 12 BU7832 10001 15000 14 BU7832 15001 20000 16 BU7832 20001 25000 18 title_id lorange hirange royalty ──────────────────────────────────────────────────────────────────────────── BU7832 20001 25000 18 BU7832 25001 30000 20 BU7832 30001 35000 22 BU7832 35001 50000 24 PS1372 0 10000 10 PS1372 10001 20000 12 PS1372 20001 30000 14 PS1372 30001 40000 16 PS1372 40001 50000 18 ──────────────────────────────────────────────────────────────────────────── Table authors ─ Structure au_id au_lname au_fname phone address city ───────────────────────────────────────────────────────────────────────────── Datatype id varchar(40) varchar(20) char(12) varchar(40) varchar(2 Null not not null not null not null null null null Default ─ ─ ─ UNKNOWN1 ─ ─ Index clust, nonclust, ─ ─ ─ uniq composite ───────────────────────────────────────────────────────────────────────────── (1) The default UNKNOWN is inserted if no data is entered. (2) The rule ziprule states that the zip code must match the pattern [0-9][0-9][0-9][0-9][0-9]. Table authors ─ Contents ╓┌────────────┌───────────────┌────────────┌─────────┌─────────────┌───────── au_id au_lname au_fname phone address city au_id au_lname au_fname phone address city ───────────────────────────────────────────────────────────────────────────── 409-56-7008 Bennet Abraham 415 6223 Bateman Berkeley 658-9932 St. 213-46-8915 Green Marjorie 415 309 63rd St. Oakland 986-7020 #411 238-95-7766 Carson Cheryl 415 589 Darwin Berkeley 548-7723 Ln. 998-72-3567 Ringer Albert 801 67 Seventh Salt Lake 826-0752 Av. City 899-46-2035 Ringer Anne 801 67 Seventh Salt Lake 826-0752 Av. City 722-51-5454 DeFrance Michel 219 3 Balding Pl. Gary 547-9982 au_id au_lname au_fname phone address city 807-91-6654 Panteley Sylvia 301 1956 Rockville 946-8853 Arlington Dr. 893-72-1158 McBadden Heather 707 301 Putnam Vacaville 448-4982 724-08-9931 Stringer Dirk 415 5420 Oakland 843-2991 Telegraph Av. 274-80-9391 Straight Dick 415 5420 College Oakland 834-2919 Av. 756-30-7391 Karsen Livia 415 5720 McAuley Oakland 534-9219 St. 724-80-9391 MacFeather Stearns 415 44 Upland Oakland 354-7128 Hts. au_id au_lname au_fname phone address city 427-17-2319 Dull Ann 415 3410 Blonde Palo Alto 836-7128 St. 672-71-3249 Yokomoto Akiko 415 3 Silver Ct. Walnut 935-4228 Creek 267-41-2394 O'Leary Michael 408 22 Cleveland San Jose 286-2428 Av. #14 472-27-2349 Gringlesby Burt 707 PO Box 792 Covelo 938-6445 527-72-3246 Greene Morningstar 615 22 Graybar Nashville 297-2723 House Rd. 172-32-1176 White Johnson 408 10932 Bigge Menlo 496-7223 Rd. Park au_id au_lname au_fname phone address city 712-45-1867 del Castillo Innes 615 2286 Cram Pl. Ann Arbor 996-8275 #86 846-92-7186 Hunter Sheryl 415 3410 Blonde Palo Alto 836-7128 St. 486-29-1786 Locksley Chastity 415 18 Broadway San 585-4620 Av. Francisco 648-92-1872 Blotchet-Halls Reginald 503 55 Hillsdale Corvallis 745-6402 Bl. 341-22-1782 Smith Meander 913 10 Lawrence 843-0462 Mississippi Dr. ───────────────────────────────────────────────────────────────────────────── au_id au_lname au_fname phone address city Table titles ─ Structure ╓┌─────────┌──────────┌────────────┌───────────┌────────┌──────┌──────┌─────┌ title_id title type pub_id price advan roya yt ce lty sa s ───────────────────────────────────────────────────────────────────────────── Datatype tid varchar(80) char(12) char(4) money money int in Null not null not null not null null null null null nu Default ─ ─ UNDECIDED1 ─ ─ ─ ─ ─ Rule ─ ─ ─ ─ ─ ─ ─ ─ Trigger deltitle3 ─ ─ ─ ─ ─ ─ ─ title_id title type pub_id price advan roya yt ce lty sa s Trigger deltitle3 ─ ─ ─ ─ ─ ─ ─ Index clust, nonclust ─ ─ ─ ─ ─ ─ uniq ───────────────────────────────────────────────────────────────────────────── (1) The default UNDECIDED is inserted if no data is entered in the column. (2) The getdate() function inserts the current date as the default if no data is entered in the column. (3) The deltitle trigger prohibits deleting a title if the title_id is listed in the sales table. Table titles ─ Contents ╓┌───────┌───────────────┌─────────────┌──────┌───────┌──────────┌──────┌──── title_ title type pub_i price advance royal ytd_s id d ty ales ───────────────────────────────────────────────────────────────────────────── BU1032 The Busy business 1389 $19.99 $5000.00 10 4095 Executive's Database Guide PC1035 But Is It User popular_comp 1389 $22.95 $7000.00 16 8780 Friendly? BU2075 You Can Combat business 0736 $2.99 $10125.00 24 18722 Computer Stress! title_ title type pub_i price advance royal ytd_s id d ty ales ───────────────────────────────────────────────────────────────────────────── PS2091 Is Anger the psychology 0736 10.95 $2275.00 12 2045 Enemy? PS2106 Life Without psychology 0736 $7.00 $6000.00 10 111 Fear title_ title type pub_i price advance royal ytd_s id d ty ales ───────────────────────────────────────────────────────────────────────────── MC3021 The Gourmet mod_cook 0877 $2.99 $15000.00 24 22246 Microwave TC3218 Onions, Leeks, trad_cook 0877 $20.95 $7000.00 10 375 and Garlic: Cooking Secrets of the Mediterranean MC3026 The Psychology UNDECIDED 0877 NULL NULL NULL NULL title_ title type pub_i price advance royal ytd_s id d ty ales ───────────────────────────────────────────────────────────────────────────── MC3026 The Psychology UNDECIDED 0877 NULL NULL NULL NULL of Computer Cooking PC8888 Secrets of popular_comp 1389 $20.00 $8000.00 10 4095 Silicon Valley PS7777 Emotional psychology 0736 $7.99 $4000.00 10 3336 Security: A New Algorithm title_ title type pub_i price advance royal ytd_s id d ty ales ───────────────────────────────────────────────────────────────────────────── PS3333 Prolonged Data psychology 0736 $19.99 $2000.00 10 4072 Deprivation: Four Case Studies BU1111 Cooking with business 1389 $11.95 $5000.00 10 3876 Computers: Surreptitious Balance Sheets MC2222 Silicon Valley mod_cook 0877 $19.99 $0.00 12 2032 Gastronomic title_ title type pub_i price advance royal ytd_s id d ty ales ───────────────────────────────────────────────────────────────────────────── Gastronomic Treats TC7777 Sushi, Anyone? trad_cook 0877 $14.99 $8000.00 10 4095 title_ title type pub_i price advance royal ytd_s id d ty ales ───────────────────────────────────────────────────────────────────────────── TC4203 Fifty Years in trad_cook 0877 $11.95 $4000.00 14 15096 Buckingham Palace Kitchens BU7832 Straight Talk business 1389 $19.99 $5000.00 10 4095 About Computers PS1372 Computer psychology 0877 $21.59 $7000.00 10 375 Phobic and Non-Phobic Individuals: title_ title type pub_i price advance royal ytd_s id d ty ales ───────────────────────────────────────────────────────────────────────────── Individuals: Behavior Variations PC9999 Net Etiquette popular_comp 1389 NULL NULL NULL NULL ───────────────────────────────────────────────────────────────────────────── The pubs database contains the following database objects: Rules pub_idrule create rule pub_idrule as @pub_id in ("1389", "0736", "0877", "1622", "1756") or @pub_id like "99[0-9][0-9]" ziprule create rule ziprule as @zip like "[0-9][0-9][0-9][0-9][0-9]" Trigger deltitle create trigger deltitle on titles for delete as if (select count(*) from deleted, sales where sales.title_id = deleted.title_id) > 0 begin rollback transaction print "You can't delete a title with sales." end Stored Procedure byroyalty create procedure byroyalty percentage int as select au_id from titleauthor where titleauthor.royaltyper = percentage View titleview create view titleview as select title, au_ord, au_lname, price, ytd_sales, pub_id from authors, titles, titleauthor where authors.au_id = titleauthor.au_id and titles.title_id = titleauthor.title_id Defaults datedflt create default datedflt as getdate( ) phonedflt create default phonedflt as "UNKNOWN" typedflt create default typedflt as "UNDECIDED" Figure B.1 shows the structure of the pubs sample database. (This figure may be found in the printed book.) Appendix C TRANSACT-SQL Keywords ──────────────────────────────────────────────────────────────────────────── TRANSACT-SQL statements contain keywords listed in this appendix. Keywords cannot be used for the names of database objects, such as databases, tables, rules, defaults, and so on, but they can be used for the names of local variables and stored procedure parameter names. ADD DESC ROLLBACK ALL DISK KILL ROWCOUNT ALTER DISTINCT RULE AND DROP LIKE ANY DUMMY LINENO SAVE AS DUMP LOAD SELECT ASC SET AVG ELSE MAX SETUSER END MIN STATISTICS BEGIN ERRLVL SUM BETWEEN ERROREXIT, EXEC NONCLUSTERED BREAK EXECUTE NOT TABLE BROWSE EXISTS NULL TAPE BULK EXIT TEXTSIZE BY OFF TO FILLFACTOR OFFSETS TRAN CHECKPOINT FOR ON TRANSACTION CLUSTERED FROM ONCE TRIGGER COMMIT OR TRUNCATE COMPUTE GETDEFAULT ORDER CONFIRM GOTO OVER, PREPARE UNIQUE CONTINUE GRANT PRINT UPDATE CONTROLROW GROUP PROC USE CONVERT PROCEDURE COUNT HAVING PROCESSEXIT VALUES CREATE HOLDLOCK PUBLIC VIEW DATABASE IF RAISERROR WAITFOR DBCC IN READTEXT WHERE DEBUG INDEX RECONFIGURE WHILE DECLARE INSERT RETURN WITH DEFAULT INTO REVOKE WRITETEXT DELETE IS INDEX ────────────────────────────────────────────────────────────────────────── + string function /* */ ** A ABS mathematical function ACOS mathematical function Aggregate functions AVG COUNT COUNT(*) MAX MIN SUM Alias add remove ALTER DATABASE ALTER TABLE ASCII string function ASIN mathematical function ATAN mathematical function ATN2 mathematical function AVG aggregate function B Batch queries bcp utility program BEGIN TRANSACTION BEGIN...END binary datatype bit datatype bldmastr utility program Boolean expression BREAK Browse mode C Case sensitivity CEILING mathematical function char datatype CHAR string function Characters wildcard CHARINDEX string function CHECKPOINT Clauses COMPUTE FOR BROWSE GROUP BY HAVING ORDER BY WHERE Clustered index Column add bind default value to bind rule to unbind default value from unbind rule from COL_LENGTH system function COL_NAME system function Comments COMMIT TRANSACTION Common key COMPUTE clause Configuration variable display or change console utility program CONTINUE Control-of-flow language Conversion function COS mathematical function COT mathematical function COUNT aggregate function COUNT(*) aggregate function CREATE DATABASE CREATE DEFAULT CREATE INDEX CREATE PROCEDURE CREATE RULE CREATE TABLE CREATE TRIGGER CREATE VIEW D Data change Database device help remove set defaults storage space for Database object help rename space used by Database options dbo use only no chkpt on recovery read only select into/bulkcopy single user trunc. log on chkpt. Database Owner change Database add group add new user back up change current change default change owner copy definitions to and from file create default size display object dependency display or change options dump help load master model pubs remove group from remove user remove rename restore damaged sample select row size space on disk DATALENGTH string function Datatype bit char Datatypes binary bind default bind rule create user-defined datetime float help image int money remove user-defined smallint sysname text text/image timestamp tinyint unbind default value from unbind rule from user-defined user_type_name varbinary varchar Date functions DATEDIFF DATENAME DATEPART GETDATE DATEDIFF date function DATENAME date function DATEPART date function datetime datatype DB-LIBRARY DBCC (Database Consistency Checker) DBCC (database consistency checker) dbo use only database option DB_ID system function DB_NAME system function DECLARE Default database change Default value bind to column or datatype create help remove unbind from column or datatype defncopy utility program DEGREES mathematical function DELETE DIFFERENCE string function DISK INIT DISK REFIT DISK REINIT DROP DATABASE DROP DEFAULT DROP INDEX DROP PROCEDURE DROP RULE DROP TABLE DROP TRIGGER DROP VIEW DUMP DATABASE Dump database Dump device add help remove DUMP TRANSACTION E ELSE END Error message EXECUTE Exit from query or procedure EXP mathematical function Expressions Boolean F float datatype FLOOR mathematical function FOR BROWSE clause Foreign key Functions aggregate conversion date mathematical row aggregate scalar aggregate string system text/image vector aggregate G GETDATE date function Global variables @CPU_BUSY @ERROR @IDLE @IO_BUSY @MAX_CONNECTIONS @NESTLEVEL @PACKET_ERRORS @PACK_RECEIVED @PACK_SENT @PROCID @ROWCOUNT @TEXTSIZE @TIMETICKS @TOTAL_ERRORS @TOTAL_READ @TOTAL_WRITE @TRANCOUNT @VERSION go command GOTO label GRANT GROUP BY and HAVING clauses Group add to database change user help remove from database H Help database device database objects databases datatypes default text dump device group index joins key permissions rule text statements stored procedure text syntax trigger text user view text HOST_ID system function HOST_NAME system function I Identifiers IF...ELSE image datatype Image functions PATINDEX SET TEXTSIZE TEXTPTR TEXTVALID Index clustered create help nonclustered remove INDEX_COL system function INSERT int datatype ISNULL system function isql utility program Isql utility program J Joins help K Key common foreign help primary remove Keywords KILL L Label user-defined LOAD DATABASE Load database LOAD TRANSACTION Local variables Locks report on LOG mathematical function LOG10 mathematical function Login authorization add remove LOWER string function LTRIM string function M master database Mathematical functions ABS ACOS ASIN ATAN ATN2 CEILING COS COT DEGREES EXP FLOOR LOG LOG10 PI POWER RADIANS RAND ROUND SIGN SIN SQRT TAN MAX aggregate function Message print on screen MIN aggregate function model database money datatype N Nest statements no chkpt on recovery database option Nonclustered index Notational conventions Null values O OBJECT_ID system function OBJECT_NAME system function Operating system file copy definitions to and from database Option display or change database set ORDER BY clause P Parameters Password add or change PATINDEX text/image function Permissions assign help revoke PI mathematical function POWER mathematical function Primary key PRINT Procedure create exit remove stored system system, enter with isql Process current pubs sample database Q Query batch R RADIANS mathematical function RAISERROR RAND mathematical function read only database option READTEXT RECONFIGURE Rename database object database table REPLICATE string function Reserved words RETURN REVOKE RIGHT string function ROLLBACK TRANSACTION ROUND mathematical function Row aggregate functions Row delete from table insert to table or view remove retrieve from database RTRIM string function Rule bind to column or datatype create help remove unbind from column or datatype S SAF Sample database, pubs SAVE TRANSACTION Savepoint Scalar aggregate function Search conditions select into/bulkcopy database option SELECT Set options SET TEXTSIZE text/image function SET Setup program SETUSER Shutdown system SHUTDOWN SIGN mathematical function SIN mathematical function single user database option smallint datatype SOUNDEX string function SPACE string function Space used by database object sp_addalias sp_addgroup sp_addlogin sp_addtype sp_addumpdevice sp_adduser sp_bindefault sp_bindrule sp_changedbowner sp_changegroup sp_commonkey sp_configure sp_dboption sp_defaultdb sp_depends sp_diskdefault sp_dropalias sp_dropdevice sp_dropgroup sp_dropkey sp_droplogin sp_droptype sp_dropuser sp_foreignkey sp_help sp_helpdb sp_helpdevice sp_helpgroup sp_helpindex sp_helpjoins sp_helpkey sp_helprotect sp_helpsql sp_helptext sp_helpuser sp_lock sp_logdevice sp_monitor sp_password sp_primarykey sp_rename sp_renamedb sp_spaceused sp_unbindefault sp_unbindrule sp_who SQL Server Administration Facility (SAF) SQL Server shutdown system start sqlservr utility program Sqlservr utility program SQRT mathematical function Statements ALTER DATABASE ALTER TABLE BEGIN TRANSACTION BEGIN BREAK CHECKPOINT COMMIT TRANSACTION CONTINUE CREATE DATABASE CREATE DEFAULT CREATE INDEX CREATE PROCEDURE CREATE RULE CREATE TABLE CREATE TRIGGER CREATE VIEW DBCC DECLARE DELETE DISK INIT DISK REFIT DISK REINIT DROP DATABASE DROP DEFAULT DROP INDEX DROP PROCEDURE DROP RULE DROP TABLE DROP TRIGGER DROP VIEW DUMP DATABASE DUMP TRANSACTION ELSE END enter with isql EXECUTE GOTO GRANT help IF IF...ELSE INSERT KILL LOAD DATABASE LOAD TRANSACTION nest PRINT RAISERROR READTEXT RECONFIGURE RETURN REVOKE ROLLBACK TRANSACTION SAVE TRANSACTION SELECT SET SETUSER SHUTDOWN TRUNCATE TABLE UPDATE STATISTICS UPDATE USE WAITFOR WHILE WRITETEXT Statistics display update Stored procedure execute help remove STR string function String functions + CHAR CHARINDEX DATALENGTH DIFFERENCE LOWER LTRIM REPLICATE RIGHT RTRIM SOUNDEX SPACE STR STUFF SUBSTRING UPPER STUFF string function Subqueries SUBSTRING string function SUM aggregate function Summary value SUSER_ID system function SUSER_NAME system function Syntax help sysalternates syscolumns syscomments sysconfigures syscurconfigs sysdatabases sysdepends sysdevices sysindexes syskeys syslocks syslogins syslogs sysmessages sysname datatype sysobjects sysprocedures sysprocesses system sysprotects syssegments System functions COL_LENGTH COL_NAME DB_ID DB_NAME HOST_ID HOST_NAME INDEX_COL ISNULL OBJECT_ID OBJECT_NAME SUSER_ID SUSER_NAME USER_ID USER_NAME System procedures enter with isql execute list sp_ sp_addalias sp_addgroup sp_addlogin sp_addtype sp_addumpdevice sp_adduser sp_bindefault sp_bindrule sp_changedbowner sp_changegroup sp_commonkey sp_configure sp_dboption sp_defaultdb sp_depends sp_diskdefault sp_dropalias sp_dropdevice sp_dropgroup sp_dropkey sp_droplogin sp_droptype sp_dropuser sp_foreignkey sp_help sp_helpdb sp_helpdevice sp_helpgroup sp_helpindex sp_helpjoins sp_helpkey sp_helprotect sp_helpsql sp_helptext sp_helpuser sp_lock sp_logdevice sp_monitor sp_password sp_primarykey sp_rename sp_renamedb sp_spaceused sp_unbindefault sp_unbindrule sp_who System tables list relationship chart syscolumns syscomments sysconfigures syscurconfigs sysdatabases sysdepends sysdevices sysindexes syskeys syslaternates syslocks syslogins syslogs sysmessages sysobjects sysprocedures sysprocesses sysprotects syssegments systypes sysusages sysusers System, shutdown systypes sysusages sysusers T Table add columns browsing copy to and from operating system file create define common key define foreign key define primary key delete row from divide into groups remove rows remove rename truncate view TAN mathematical function text datatype Text functions PATINDEX SET TEXTSIZE TEXTPTR TEXTVALID Text read from database write into database TEXTPTR text/image function TEXTVALID text/image function timestamp datatype tinyint datatype TRANSACT-SQL Transaction log back up dump load place in separate database device size of database device required storage space Transaction begin end rollback save Trigger create help remove trunc. log on chkpt. database option TRUNCATE TABLE U UPDATE STATISTICS UPDATE UPPER string function USE User add login authorization add new change group current help impersonate another remove from database remove login authorization USER_ID system function USER_NAME system function user_type_name datatype Utility programs bcp bldmastr console defncopy isql sqlservr V Value null summary varbinary datatype varchar datatype Variables global. See Global variables local Vector aggregate function Views create define common key define foreign key define primary key remove W WAITFOR WHERE clause WHILE Wildcard characters WRITETEXT