Statspack better than sql-trace? (was: High elapsed/cpu ratio)

From: Jan Krueger <jk_at_stud.uni-hannover.de>
Date: Tue, 11 Sep 2007 21:14:50 +0200
Message-ID: <46e6e89d$0$29374$4c56b896@news-read1.lambdanet.net>

David Fizjarell wrote:
> Level 8 gives you the wait statistics, level 12 adds to that bind
> variables, if I remember correctly, so you should have the wait
> statistics necessary to troubleshoot this query. Statspack might
> provide better information as it can show waits other than those
> already in the 10046 trace.

Is this really true? I would say, waits not showing up in the tracefile don't matter at all for the statement you want to optimize.

Jan Received on Tue Sep 11 2007 - 14:14:50 CDT