Re: Oracle Myths

From: William Rice <ricew_at_operamail.com>
Date: 23 May 2002 15:15:49 -0700
Message-ID: <1f1a539b.0205231415.51b83543@posting.google.com>

"Niall Litchfield" <n-litchfield_at_audit-commission.gov.uk> wrote in message news:<3ce36ec0$0$8513$ed9e5944_at_reading.news.pipex.net>... <SNIP>
> >
> > Seperate tables and indexes for performance reasons.
>
> see the huge thread on this earlier at http://shrinkalink.com/201
>
> I think my summary of this is that in general seperating data and indexes
> should be done for
> 1) management reasons only.

While I didnt read the whole thread, I guess what I got from it was the old statement, benchmark your system and find out...

<SNIP>

An example of where separating indexes from a table might be beneficial would be if you happen to do lots of full table scans on a particular table. If you make it the only inhabitant of a tablespace (or tablespaces if it is partitioned), you avoid having to skip past all of the extents you are not interested in. I am not saying there are no down sides to this solution, but I think being able to fragment index _can_ give performance benefits in some situations.

Will
P.S. I am a bit new to the Oracle arena, but have been doing DBMS management for an undisclosed product for awhile :) P.P.S. This is assuming raw devices are used for the tablespaces Received on Thu May 23 2002 - 17:15:49 CDT