RE: Index compression on Oracle 9.2

My apologies. Perhaps I misinterpreted the content at http://www.jlcomp.demon.co.uk/faq/compress_ind.html

Rich

-----Original Message-----
From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Jurijs Velikanovs Sent: Tuesday, January 31, 2006 12:43 PM Cc: oracle-l_at_freelists.org
Subject: Re: Index compression on Oracle 9.2

> plus there'll be obvious CPU usage because it'll need to uncompress
for
> every read and compress for every write. The few GBs of storage to be
saved
I wouldn't say that CPU usage increasing is so obvious. Oracle actually doesn't do any compression
(oracle always toyed to avoid compression algorithms usage as Zip and other classical utilities are using).
In an "Index compression" feature case Oracle avoids duplicate records in leaf blocks. I would call the feature differently than Oracle did, because the name doesn't reflect the meaning of functionality.

By my opinion the feature is absolutely grate (whilst table compression is rubbish). If you are using it in right place it will not just save you a space on disk but reduce pressure on Memory and CPU. We have used it with grate success in different projects.

I would recommend you to test it on your test database. Then you would get an impression how many memory it will save you.

Jurijs

--
http://www.freelists.org/webpage/oracle-l