Via va Viaindex

[Cousins, Michael]

Flexibility and efficiency are two highly touted reasons for using an RDBMS.
Make the rdbms do the work.

If the query writer specifies ordering on, coincidentally, on the same
segments as an existing index then chances are that index will likely be the
chosen retrieval method. However, if all of the conditions for the retrieval
match the use of a different index, such as the cardinality of the combined
segments presents less uniqueness than an existing unique index -  all other
considerations equal - then the other index may be used. Everything depends
on the content of the query.

Single table queries are far less sophisticated than queries spanning
multiple tables and several layers of subqueries; however, the purpose of
the optimizer remains the same: to consistently choose the most efficient
method for retrieving the data based upon the query request. It may not
always make what seems is the best decision, but those are rdbms engine
issues - not application code issues - and there are query tuning procedures
available. 

If a Quiz report, or native sql query for that matter, was created several
months in the past - the nature and volume of data in the related tables may
be dynamic enough, and may have changed significantly enough, so as to
invalidate the retrieval path specified in the code. I'd rather have the
query optimizer evaluate my data and chose the best path than have to
maintain the application code based upon patterns of data.

There's a lot more to this issue than this simplistic explanation. Suffice
to say that I would be interested to see an example of an rdbms that allows
the query-writer, in whatever client, to over-ride the query optimizer - and
see the reasons behind having an optimizer and arbitrarily being able to
turn it off.
Michael Cousins
Senior Technical Specialist, Alliance Team,
Worldwide Customer Support
Cognos, Inc.

= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Subscribe: "subscribe powerh-l" in message body to majordomo@lists.swau.edu
Unsubscribe: "unsubscribe powerh-l" in message to majordomo@lists.swau.edu
powerh-l@lists.swau.edu is gatewayed one-way to bit.listserv.powerh-l
This list is closed, thus to post to the list, you must be a subscriber.