Re: Extent Management 101

From: Anurag Varma <avdbi_at_hotmail.com>
Date: 13 Aug 2004 08:07:38 -0700
Message-ID: <7171ca2d.0408130707.714e5d74@posting.google.com>

"Christian Antognini" <christian.antognini_at_trivadis.com> wrote in message news:<411b3212$1_at_post.usenet.com>...
> **** Post for FREE via your newsreader at post.usenet.com ****
>
> Hi Anurag
>
> > I'm not entirely sure .. but even the alter table .. minimize
> records_per_block
> > might fix the problem too..
>
> Minimize records per block just fix a limit to future inserts. No
> modification of the segment's storage is performed.
>
> Chris
>
>

--snip--

Er .. you seem to be missing the point. minimize records_per_block will
calculate the *Actual* maximum number of records that can be placed in a block of that table (See Metalink Note: 103490.1). This enables the bitmap creation
later on to use the *Actual* max value instead of guessing on the value.

Now the problem of "table too fragmented" error is being caused because the guessed value is wrong. Thats why I state that if the *actual* value is known, then this error should not be thrown.

And also see the following metalink thread: http://metalink.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=FOR&p_id=165572.995

wherein this exact solution was followed by a poster and it worked.

Do you agree?

Anurag Varma Received on Fri Aug 13 2004 - 10:07:38 CDT