CP/M/CPM_CDROM.iso/cpm/database/filebase.lbr/FILEBASE.DZC/FBASE.DOC

Text File | 1990-03-18 | 25.0 KB | 502 lines

                   A BRIEF INTRODUCTION TO FILEBASE

This shareware version of FILEBASE will run without installation or
modification on any CP/M 2.2 or later system.  However, the screen will be
cleared by scrolling the old information off the top of the screen.  A much
more enhanced instantaneous clear-screen method can be implemented but it
is dependent on codes required by your hardware.

ESTABLISHING THE CLEAR-SCREEN CODES FOR YOUR COMPUTER

The clear-screen codes that FILEBASE uses are external to the executable
program code.  These codes are to reside in a file called FBCLRSCR.CDE
which can be created with FILEBASE.

Create a one field file by the name of FBCLRSCR.CDE with FILEBASE
Option 6 - Create a File.  Enter into the first record the number of
codes that are to follow in subsequent records.  There should be one
code per record as shown in the following example.

Many CP/M computers such as Kaypro, Epson, Xerox, Morrow, Osborne,
some Televideo's use the decimal value 26 as the clear-screen code.
This number is also applicable if your computer documentation tells
you that your terminal (CRT) emulates an ADM-3A.  For these, the
FBCLRSCR.CDE file should contain the numbers:
                     1
                    26

The following sequence works with any ANSI standard terminal.
                    4  (number of codes to follow)
                   27
                   91
                   50
                   74

With some computers, the terminal can be optionally defined to conform
to the ANSI standard.

If the codes you use are not correct, you will notice when running FILEBASE
that the screen is not properly cleared.  Either obtain the correct
numbers or delete the FBCLRSCR.CDE file and FILEBASE will revert to the
scrolling method.

Registered users can call EWDP for the correct codes for each computer.
Upon registration, the clear-screen code file specific for the computer
will be included in the update disks sent.


SAMPLE FILES

Sample data files and corresponding prompt files to be used for
familiarization with program functions are:

Data file                Prompt File               Comment

INFILE1                  INFILE1.PMT               Names & addresses
INFILE2                  INFILE2.PMT               Names & addresses
LISTC.FXD                LISTC.PMT                 Names & addresses
CHECKS.DTA               CHECKS.PMT                Check Register data
ARTICLES.FXD             ARTICLES.PMT              Article references

The contents of the data files may not correspond exactly field
for field with the printed versions used in the Tutorial Section
of the Manual.  They may vary for illustrative purposes.

INFILE1 and INFILE2 can be used with all menu opitions with the exception
of Option 9 - Update.   Option 9 requires a one-time conversion to
fixed length records.  This can be performed with Option 10 or as
part of a sort or select operation with Options 1, 2, or 3.

The two files can be merged into one because they have the same
number of fields and the same kind of logical information in each
of the corresponding fields.  That is, field one is a fullname in
each file, field 2 is a last name, etc. etc. and field 6 is a
numeric quantity which the prompt file defines as CONTRIBUTION.

LISTC.FXD is identical in format to the first 2 except that it has
already been converted to fixed length records and can be used
immediately with Opition 9 for random access, search, and to
change information such as a new address.

CHECKS.DTA is also a fixed record length file and is a typical
check register layout.  It can also be used to illustrate the
14-column summary profit and loss statement generated by the
PROFLOSS program.  EXPENSE.CAT is the expense categories
file as desribed on manual pages 72-74.

ARTICLES.FXD is a file suitable to serve as an example of print
capabilities on longer text fields such as TITLE and SYNOPIS
within this sample.  The LFS (LABEL FORMAT TO SCREEN) and
TBF (TABULAR BLOCK FORMAT) within Option 4 would be illustrative
of the word-wrap capabilites in these report styles.


DESCRIPTION

FILEBASE files consist of "comma delimited field records".  A record
is a group of related items of information (fields)  - one person's
'name,street,city,phone' would be a record with 4 fields.  Each item
of information (name) is automatically separated from the next
(street) by a special comma separator.  You will not see the commas or
be concerned about them.  The importance of this method is that items
of information can vary in length within each field and from record to
record.  Most other data managers separate fields from each other by
setting aside a certain number of character positions for each field.

These files can be used directly with most letter-merge software such
as MailMerge, Multimate, Word Perfect, Word, and others. Even pre-
existing comma delimited field record files can be processed directly
no matter how they were created. For example, dBase-II can make a copy
of its fixed length field files in comma delimited form.  Pre-existing
block format field records can be converted to comma delimited.


FILE COMPATIBILITY

There is high degree of compatibility with programs written in BASIC.
This means that if you program in BASIC, the files that were created
with FILEBASE can be used by your programs.  The data manager will
perform all the necessary processing so that you are not burdened with
writing input, update, sort, select, and other manipulation routines.


APPLICATIONS

FILEBASE is ideal for mailing lists, literature citations, membership
data, client lists, parts lists, keyword information and a host of
other applications because it can create and process items of
information even if they vary considerably in length from record to
record.  You do not have to worry about how much space to allot for
every item of information as you do with other data managers. You can
enter long lines of information (up to 255 characters per field) and
output such information without it running off the paper or looking
odd on the screen.

FILEBASE can be used for mailing lists with great flexibility.  Lists
with up to 40 items of information in each record can be produced and
any item can be selected for printing labels or envelopes, or line
listings including those that may have non-address type information
such as "identification number", "chapter & year", etc.    FILEBASE
even sorts on the last name part of a full name field or a five digit
zip code part of the address line.  There are features that print any
number of lines in 'label format' to use for bins, boxes etc.


FILEBASE VS. FIXED LENGTH FIELD DATA MANAGERS

There is an important difference between FILEBASE and classical data
management systems.  Other systems require the user to pre-define a
formal structure for a file such as number of fields, their type and
length. Unless the definition has been very well pre-defined, some
field lengths may prove excessive (wasting disk space) and others too
small not allowing all data to be entered.

FILEBASE permits field lengths to vary and maintains random access
capability by means of a fixed total record length which can easily be
modified at any time without affecting individual fields. A formal
file structure is not necessary which makes it very easy to set up a
file.  You need not figure out in advance how long each field has to
be.  FILEBASE actually helps you to decide on an optimal record length
after you enter records.

When you create a file, records keyed in are variable length until
converted to fixed length which is required to use one particular
FILEBASE option.  After you have made a reasonable group of entries
into the file and decide to convert to fixed lengths,  the program
will tell you the length of the longest record up to that time.  This
will help you to decide on a proper fixed record length.

Even if, after converting to fixed length, you come across an
unusually long entry, you can easily handle it by changing to a longer
record length.  Once you convert to fixed length, or re-convert to a
longer length, the program will automatically keep new records the
same by padding them with blanks on the right side of the record. You
need not be concerned about the length of data you put into fields.


FEATURES

Files can be sorted, or sort/merged into one or separated into logical
subsets by various methods.  Records can be located by a host of menu
driven methods including the familiar logical comparators (EQ, GT,
etc.), by searching for words or parts of words, or by record number
methods.  Fields can be dropped, rearranged or new ones added.

Records can be printed individually or in logical groups continuously
or with pause for paper or label or envelope change. Format it to any
size label, index card or on whatever you can put into the printer.
Print line listings or fields under each other (sort of label format).
Print with or without field prompts, with or without record numbers
and add keyboarded comments to be printed.  Change the information in
the fields and print the record at the same time, any number of times,
in any number of different formats to get what you want at the time.

There is comprehensive error trapping with immediate feedback for
unacceptable responses to prompts.  Presence or absence of files is
verified and existing files are prevented from being accidentally
overwritten.


STARTING FILEBASE

With the FILEBASE disk in the logged-on drive, just type FILEBASE,
select an option from the menu, and answer the ensuing prompts.

Only those prompts are presented which relate directly to the
currently selected menu choice.  Their sequence follows a logical
pattern which is easy to follow and understand.  Most functions of
FILEBASE can be used without reference to the manual.

However, the manual includes a step-by-step tutorial that illustrates
each of the prompts in their order of occurrence.  It is intended to
be used during actual processing of either your files or the practice
files that are on the FILEBASE disk.  All possible responses are
outlined together with an explanation of the ramifications of each
valid choice.  In addition to a discussion of what you can do, there
is frequently an explanation of how FILEBASE will handle your
instructions.

The tutorial follows the logic of the program because it was prepared
using actual portions of the source code.  Whenever the program
branches to a subroutine, so does the explanation.  In fact, the
actual source code was word processed to extract the prompts and
convert them to bold print and to additionally flag them with an
asterisk.

FILEBASE is not only menu and prompt driven but it has  comprehensive
error trapping. For example, invalid characters in user entered file
names are located at the time that the name is entered.  Existence of
input files is verified and output file names are checked to guard
against existing ones from being overwritten.  Input file names are
checked against output file names and inappropriate entries result in
immediate feedback.  All replies to prompts are verified for
acceptability as soon as they are entered.  Repetitive operations use
a single keystroke (return not required). Lower or upper case can be
used in reply to most prompts.

If you use an existing file, FILEBASE will find out how many fields
there are in the records. It is not necessary for you to remember the
number of fields or their sequence before you start. FILEBASE will
display as many records as desired with each field numbered so that
you can decide what fields to use for the various processing options.
If you decide to merge two files, FILEBASE will tell you if the number
of fields in the two files are the same.  It will tell you whether the
file you are processing is comprised of fixed length records or
variable length ones, how many records there are, and the length of
the longest record.


CREATING A NEW FILE

You can create a new file, decide on how many fields (items of
information) it will have (up to 40), and assign names to be displayed
as prompts for date entry.   Designate a file name. Blank fields are
displayed so that you can assign prompts (field names) to be used for
data entry. During the data entry process, the prompts assist in
entering information.  See page D-14 under GLOBALS for codes you can
assign to represent repetitive items of information.

When you are through entering data, prompts can be stored under any
file name you assign. To differentiate prompt files from data files,
one can use the name of the data file but have the suffix something
like .PMT.  Prompts can be recalled for use in appending data to this
or any other file that has the same fields in the same sequence.

When you create a file and key in records, they will be variable
length ones.  Each will be only as long as the characters you keyed
(plus comma delimiters added by the program). After keying in a group
of records, FILEBASE will be able to tell you the length of the
longest record up to that point and allow you to make and informed
judgement on what length to use when converting to fixed.  You do not
have to decide on individual field lengths.

The general concept of maintaining files is that once the file has
been converted to fixed record lengths, it becomes the master file
with the records kept in what is referred to as "arrival sequence".
If a sorted report is needed, the file is sorted to an output file to
be used for the report.  To save disk space, one can output variable
length records.  The sorted file can be held until new records which
are added to master file make the sorted file obsolete.

The process of creating a new file is so simple (because a formal
structure with field lengths and types is not needed), that one can
easily setup temporary files to later be merged with more permanent
ones.  For example, when new FILEBASE users are entered into a file,
it is a new one created weekly to be merged later into the master.

Creating a new file is so easy that it even is worth while to create a
new file for just one record.  How about the name and address of your
mortgage company if they do not provide self addressed envelopes.
Enter the one address (or others if applicable), then use Option 4 to
print 12 copies of each record on business envelopes.  Give the tab
position as 48, number of lines to skip as about 10, and insert the
envelope to the desired depth.  The address will print and eject
because of the 10 line feeds. You can have a years worth of envelopes
for each address entered in less than 5 minutes.

The file you create is not cast in concrete so you do not have to plan
for every eventuality.  Later, if you discover a new field is needed,
add it with Option 5, fill it in with data with Option 8 or 9 and you
can also move it to a different spot in the record with Option 8 and
at the same time drop fields you no longer need.


CONVERSION TO FIXED LENGTH RECORDS - Fields Remain Variable Length

OPTION 10 is a direct method for converting variable length records to
fixed for use with OPTION 9-UPDATE & SEARCH.  Options 1, 2 or 3 can
also output fixed length, however these 3 can also convert fixed back
to variable as well.  The latter could be useful if, after having
converted to fixed length you find that a shorter length would
suffice.  You first convert your fixed length records back to variable
and then convert back to a shorter fixed length.  If you have no
reason to sort, use OPTION 2 because it will be faster.

If select/exclude testing is not required and you wish to output the
entire file of  records,  specify a test that is sure not to exclude
any records but will serve merely to change record lengths. (Ex. Test
against a LAST NAME field and EXCLUDE if EQ 'ZZZZZZ').


UPDATING RECORDS

If you've processed a file using options 1, 2 or 3 or 10, and output
fixed length records, you can update or search the information in that
file with random access by record number or by key value.

The file must contain at least 2 records for FILEBASE to insure that
it is indeed a fixed length records file.  With only one record the
program assumes that the file is a variable length one and will return
to the main menu.

FILEBASE will display the contents of the fields numbered so that you
can assign prompts to assist you in altering field contents.  Prompts
already stored in a file from a prior run, can be recalled for use.
You can select one field at a time (ex. last name) to produce an index
to access records by means key value entry.

Retrieved records can be printed and any of the field contents
changed.  Choices for printing include TAB position, number of lines
to skip between each  and addition of a keyboarded comment.

Accessing records can begin as soon as the prompts are available
either by recall from a file or by assignment.  The update process is
by record number or by key.  If by number, you enter a relative record
number.  That record is retrieved from its position in the file and
displayed for changes or printing.  If by key, you get a menu of
comparators followed by a prompt to enter a key value.  The choice of
comparators will be the usual ones (Greater Than, Less Than, Equal To,
Between, etc.) and will include two for string searching.

When changing field contents, FILEBASE will insure that the sum total
of the altered field contents plus delimiters plus quotes is not
greater than the length of the other records in your fixed length
file. If the record is too long, you will be advised of how much to
shorten it before it can be written to the file.  You could abbreviate
the contents of any of the fields to accommodate to the maximum length
of the record or you could convert to longer record length.

If you know in advance that the updating will cause records to be
longer than the present fixed length will permit, you can first use
OPTION 10, 1, 2, or 3, any of which will allow you to duplicate your
file specifying a greater record length.  If you have no reason to
sort or select or exclude records, the fastest conversion process is
with Option 10.

With fixed length records files, you can move directly to Option 7 -
APPEND, without returning to the main menu.


MORE ON CONVERTING TO FIXED LENGTH RECORDS WITH OPTION 10

FILEBASE  converts variable length records to fixed length records by
padding blanks to the right side of the record so that FILEBASE option
9 can utilize the file for random access.

Random access is a method to retrieve a record from its position by
going directly to it as opposed to starting at the beginning of the
file and reading records until the right one is reached.  If all the
records in a given file are the same length, the operating system of
the computer combined with the disk controller can calculate any
record's physical position by means of a relative record number.

FILEBASE implementation of this concept is different from most data
management systems in that you do not have to pre-define field lengths
by a "best guesstimate" of how wide they should be to accommodate the
longest entry that might ever be made into that field.  Unless these
lengths have been very well pre-defined, some may prove excessive
(wasting disk space) and others too small (restricting the number of
characters that may be entered).

FILEBASE permits field lengths to vary and maintains random access
capability by means of a fixed total record length which can easily be
modified at any time without affecting individual fields.  You do not
have to sit down and try to figure out in advance how long each field
has to be. FILEBASE actually helps you to decide on an optimal record
length after you have entered records.

When you create a file, records keyed in are maintained as  variable
length ones until you decide to convert them to fixed length.   After
you have made a reasonable group of entries, you can use this option
to convert the records to fixed length.  The program will report the
length of the longest record to help you to decide on a total record
length.   After converting to fixed length, even if you come across an
unusually long entry, you can easily handle it by using this option to
change to a somewhat longer record length.  Once you convert to fixed
length, or re-convert to a longer length, the program will
automatically keep newly added records the same by padding them with
blanks on the right side of the record. You never have to be concerned
about the length of data you put into individual fields.

Option 10 will tell you the length of the longest record and  will
permit you to select any length equal to or greater than the  longest.
Your  choice  will depend on how confident you are that the number  of
records  entered  is  sufficient  to  base  your  judgement   on.    A
comfortable  decision would be to ask for a length that is about 10 or
20  longer than the longest record.

Records will be converted to fixed length and output to a new file.
The input file many already be fixed length and this option can be
used to increase its length, however the length cannot be decreased
unless you first remove the blanks with Options 1, 2 or 3.


RECORD COUNTING AND RECORD LENGTH CHECKS

FILEBASE can process comma delimited field records (Mailmerge and
BASIC format) without need to create a special structure.  Also,
FILEBASE created files have no header records as do some other data
managers. This is to keep the files directly compatible with
MailMerge, MultiMate, Word and other word processors with letter-merge
capabilities, as well as with dialects of BASIC.

With most data managers there is a header record or structure file
keeps track of the record length, number of fields, and number of
records in a given file.  Data managers that use header records can
immediately determine the properties of the file by reading the
header.

Some main menu options require checking the properties of the file.
Since FILEBASE files do not have header records, the properties of the
file must be learned by direct examination.  For some options, this
requires checking record lengths and counting them.


RECORD DELETION - General

Most data managers delete records by marking them with a special
delete marker.  The record is not physically removed from its position
in the file but the data manager recognizes the marker and ignores the
record when printing reports and other processes.  Some programs
automatically re-use deleted records by writing newly entered ones
over the deleted ones.  Others, physically remove the records with a
process that may be called 'file maintenance' or 'pack operation.'

Such processes actually copy the file to a new file name, leaving out
records with the delete marker.  The old file is then automatically
deleted and the new one renamed with the original file name.  This
process tightens up the file by eliminating the existence of 'deleted'
records and may be totally transparent to the user.

Some word processors also function in this manner.  In the case of
WordStar, the old file is renamed with the file type .BAK and the
newly re-written file receives the original file name.


Record Deletion with FILEBASE

Because the FILEBASE concept of data management requires that files be
in pure data format so they can be used without modification by other
programs, embedding delete markers to records is not provided for.
While delete markers are recognized by the data manager using them,
they are totally foreign to other programs and may interfere with the
proper use of the files.

Deleting records from FILEBASE files can be performed by one of two
methods.  For an occasional delete in a fixed length records file, one
can simply access any record no longer needed and filling in the
fields with information from any new entry that is to be made.

However, the principal method would be to use Option 2 to exclude them
from a newly written file.  This process is analogous to a 'file
maintenance' step or 'pack operation' with other programs.  With
FILEBASE it is a conscious and observable process.

The exclude process with Option 2 can be based on the LS-List of
Record Number technique or by changing some field (in the records to be
deleted) to some unique value.  For example, the first field of each
record to be deleted can be changed to the word DELETE.  Using the
Field Test method within Option 2, EXCLUDE all records EQ DELETE from
the new file.   If MYFILE has records with the first field bearing the
word DELETE, and MYFILE2 is named as the output file, do a test on
field 1 and EXCLUDE if EQ DELETE.  To confirm that only deleted
records were truly excluded, name a rejected records file.  All
excluded records will be written to that file.  Look at that file for
confirmation.

With the LS-List of Record Numbers technique, first identify the
record numbers to be deleted by using Option 4 and then enter them
into the exclude list.  See the manual for the LS technique.