home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.cs.arizona.edu
/
ftp.cs.arizona.edu.tar
/
ftp.cs.arizona.edu
/
tsql
/
doc
/
tsql.mail
/
000045_csj@iesd.auc.dk _Sun Mar 21 16:04:47 1993.msg
< prev
next >
Wrap
Internet Message Format
|
1996-01-31
|
3KB
Received: from iesd.auc.dk by optima.cs.arizona.edu (5.65c/15) via SMTP
id AA21740; Sun, 21 Mar 1993 08:04:48 MST
Received: from yellow.iesd.auc.dk by iesd.auc.dk with SMTP id AA26193
(5.65c8/IDA-1.5/MD for <tsql@cs.arizona.edu>); Sun, 21 Mar 1993 16:04:47 +0100
Date: Sun, 21 Mar 1993 16:04:47 +0100
From: "Christian S. Jensen" <csj@iesd.auc.dk>
Message-Id: <199303211504.AA26193@iesd.auc.dk>
To: tsql@cs.arizona.edu
Subject: Benchmark schema
The TSQL Benchmark Draft - TASK I
*********************************
Below, I briefly comment on the topics that have been brought up so
far. I a few days, I will attempt to update the benchmark document to
reflect the recent discussions while maintaining consistency. I will
also include a straw proposal for data for the benchmark schema.
1. Scope of the initial version of the benchmark.
Two extensions of the scope have been suggested, namely
(a) it should be possible to use the same relation more than
once in each query
(b) queries involving aggregation facilities should be included
It has also been argued that these extensions should wait until the
next version. This is the current, unchallenged proposal.
Aggregation is very important, but is also a substantial extension.
For that reason, I would like to postpone (b) to the next version.
I would like to propose another extension:
(c) it should be possible to ask queries involving valid time
and user-defined time
Queries like "Who made more than 30K before turning 25" are very
natural, and a well-designed query language should be able to express
such queries conveniently.
The current schema has no user-defined time attributes. Such
attributes are needed in order to support (c). The most obvious
extension appears to be to add birth day as an attribute in the Emp
schema.
If (a) and (b) are to be postponed to the next version, it is
reasonable to also postpone (c). I propose that the introduction is
divided into two. One part will describe the goal of the benchmark.
The second part will describe the scope of the current version.
Limitations (a) - (c) should be indicted clearly. In addition, the
second par will reflect our discussions about how the scope should be
extended in the next version.
3. Additional attributes.
The inclusion of additional attributes has been argued.
Support for (c) above motivates the inclusion of a user-defined time
attribute (birth day, to be added to Emp). This addition can wait
until (c) is included.
Inclusion of time-invariant (same as "non-temporal") attributes has
been proposed. It is my feeling that the motivation for this is to
obtain a schema better suited for queries involving aggregates (?). A
attribute like "gender" should then be added when aggregation is
introduced. I propose that this addition is postponed until (b) is
included.
Inclusion of a multivalued dependency which is not a functional
dependency has been proposed, possibly in a future version. This may
be accomplished by adding a "skills" attribute to the Emp schema. I
propose that we maintain a list of useful extensions like this one.
Best regards,
Christian S. Jensen
Aalborg University