Re: NL & Hash Outer Join question

From: <jonathan_at_jlcomp.demon.co.uk>
Date: Fri, 26 Jan 2007 11:11:14 +0000
Message-Id: <E1HAOzC-0001XM-EE@pr-webmail-2.demon.net>

Could be something changing in a different part of the plan which changes the cardinality estimates of the driving table.

Could be a change in predicate generation.

Why not run a 10053 trace on both versions and compare the two sets of parameters reported for differences.

How about posting the plans (from a dbms_xplan output - with predicate information).

Regards
Jonathan Lewis

> Date: Thu, 25 Jan 2007 15:59:06 -0500
> From: rjamya <rjamya_at_gmail.com>
> Subject: NL & Hash Outer Join question
>
> What might cause a query using NL (in 9204) use Hash Outer Join (in 9208)
> when all optimizer and memory parameter being same. I checked optimizer
> parameter values using 10053 trace.
> The tables involved are analyzed identical, contain identical data (9208 was
> cloned from 9204), they run on identical servers and only one person logs in
> and runs the test. Memory sizes etc are all identical.
>
> Any ideas?
> Rjamya
> ----------------------------------------------
> Got RAC?

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

Received on Fri Jan 26 2007 - 05:11:14 CST