Re: Wrong optimizer estimate ?
From: Greg Rahn <greg_at_structureddata.org>
Date: Mon, 8 Nov 2010 20:35:23 -0800
Message-ID: <AANLkTi=Y-XH5LW7iPwqsTbwSpy+cNEK+UyCCdHoBRdfa_at_mail.gmail.com>
You really only need to worry about deltas in cardinality estimates vs actuals when they are off enough to change the execution plan. In this case both plans are identical.
Date: Mon, 8 Nov 2010 20:35:23 -0800
Message-ID: <AANLkTi=Y-XH5LW7iPwqsTbwSpy+cNEK+UyCCdHoBRdfa_at_mail.gmail.com>
You really only need to worry about deltas in cardinality estimates vs actuals when they are off enough to change the execution plan. In this case both plans are identical.
Representative stats generally results in good plans, however, non-representative stats do not always result in bad plans.
On Sun, Nov 7, 2010 at 1:53 AM, Michael Seiwert <Michael.Seiwert_at_gmx.de>wrote:
>
> Fortunately there is no performance problem with the query yet, but I've
> learned that even small differences in data combined with small
> miscalculations may have a huge impact on the query perfomance in the
> future.
>
-- Regards, Greg Rahn http://structureddata.org -- http://www.freelists.org/webpage/oracle-lReceived on Mon Nov 08 2010 - 22:35:23 CST