Re: Merge statement not using index
From: Sandra Becker <sbecker6925_at_gmail.com>
Date: Mon, 26 Mar 2012 07:24:01 -0600
Message-ID: <CAJzM94BMq-ihi0V0Ph6sgMRwirztgmMX_tYVAWs=CV=+zQ7pTQ_at_mail.gmail.com>
It is a merge. Tables are far from small which is why an FTS is problematic. It does not use binds, no need for them. It is ignoring all index hints. All indexes are valid. Shared pool is not undersized. The only predicate is select everything from the source where the document_key is in the lookup table. The only thing I haven't done is regather stats on the target for awhile. Stats are gathered on the source every month. I'll try again after getting new stats. Thanks for the input. Sandy
Date: Mon, 26 Mar 2012 07:24:01 -0600
Message-ID: <CAJzM94BMq-ihi0V0Ph6sgMRwirztgmMX_tYVAWs=CV=+zQ7pTQ_at_mail.gmail.com>
It is a merge. Tables are far from small which is why an FTS is problematic. It does not use binds, no need for them. It is ignoring all index hints. All indexes are valid. Shared pool is not undersized. The only predicate is select everything from the source where the document_key is in the lookup table. The only thing I haven't done is regather stats on the target for awhile. Stats are gathered on the source every month. I'll try again after getting new stats. Thanks for the input. Sandy
On Sun, Mar 25, 2012 at 1:44 AM, Sidney Chen <huanshengchen_at_gmail.com>wrote:
> To make it clear, better provide the execution plans with predicate info
> and peeked binds value.
>
>
> --
> Regards
> Sidney Chen
>
>
>
-- Sandy Transzap, Inc. -- http://www.freelists.org/webpage/oracle-lReceived on Mon Mar 26 2012 - 08:24:01 CDT