Re: Performance Tuning

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Fri, 8 Oct 1999 18:32:06 +0100
Message-ID: <939403957.17225.0.nnrp-12.9e984b29@news.demon.co.uk>

Looking at the

>table scans (long tables)

70460

>table scan blocks gotten

126911699

>table scan rows gotten

1.257E+09

everyone on this newsgroup is likely to suggest that you have an application code problem.

select executions, disk_reads, buffer_gets, sorts, sql_text from v$sql where disk_reads > 10000

might be a good starting point.

Where disk-reads is close to buffer gets you have tablescans - identify the code, see if you can determine where it came from, and if there is an efficient access path for it (e.g. by creating needed indices, getting rid of functions applied to columns etc.)

Jonathan Lewis
Yet another Oracle-related web site: http://www.jlcomp.demon.co.uk

Stephen Gabor wrote in message <37fe2501_4_at_news1.prserv.net>... > Received on Fri Oct 08 1999 - 12:32:06 CDT