Re: Force specific plan to be used

From: Chris Taylor <christopherdtaylor1994_at_gmail.com>
Date: Thu, 31 Oct 2019 21:39:08 -0500
Message-ID: <CAP79kiQNE1954b4fT1767t2_gLsxECXXtUEyA047rscm2eiTQg_at_mail.gmail.com>



But wouldn't that show up in the plan? I must be missing something.

How can you tell by examining the predicate section where the filtering is occurring?

Chris

On Thu, Oct 31, 2019, 8:35 PM Tanel Poder <tanel_at_tanelpoder.com> wrote:

> Even if the plan hash values are the same, go still ahead and compare the
> predicate sections of the good vs bad child cursors.
>
> Predicate existence (or placement) is not part of the plan hash value - so
> how early you're filtering the rows may differ.
>
> Tanel Poder
> https://blog.tanelpoder.com/seminar
>
> On Thu, Oct 31, 2019 at 2:03 PM Jeffrey Beckstrom <jbeckstrom_at_gcrta.org>
> wrote:
>
>> I have a couple of SQL statements that have multiple child cursors. Each
>> child cursor has the same plan hash value. The plans all show "this is an
>> adaptive plan (rows marked '-' are inactive)". The difference is that on
>> the "good" child cursor, the plan also shows "statistics feedback used for
>> this statement". Since all of the plans have the same plan hash value, I
>> can not use baselines (or can I).
>>
>> Any suggestions on how to force Oracle to always use the "good" child
>> cursor.
>>
>> Jeffrey Beckstrom
>> Lead Database Administrator
>> Information Technology Department
>> Greater Cleveland Regional Transit Authority
>> 1240 W. 6th Street
>> Cleveland, Ohio 44113
>>
>>

--
http://www.freelists.org/webpage/oracle-l
Received on Fri Nov 01 2019 - 03:39:08 CET

Original text of this message