Re: Incorrect Migrated/Chained rows...

From: Howard J. Rogers <hjr_at_dizwell.com>
Date: Wed, 28 Apr 2004 07:00:18 +1000
Message-ID: <408ec9db$0$436$afc38c87@news.optusnet.com.au>

MAK wrote:

> Mr. HJR,
>
> Thanks for all your suggestions and recommendations reg. pctfree and
> unrecoverable. But that was not something I was looking for.
>
>
> If you have read my post carefully, I was looking for an explaination
> as to why I have migrated rows when there was no "UPDATE" activity ( I
> just created table using CTAS and ran analyze to count
> chained/migrated rows). I also mentioned that there was no row that
> had row length more than the block size so chaining of rows can be
> ruled out.
>
> Can some one explain this?

If *you* read *my* post carefully, you will hopefully get the idea that anyone who sets PCTFREE to something stupid and uses a piece of syntax that has been deprecated for more than 8 Oracle releases is just asking for weird things to happen. You might not have been "looking for" such comments, but they are true nonetheless and relate to your original question.

Explanation? Who knows. Maybe it's a transaction slot issue.

Maybe it's a bug.

Does it really matter? Delete the rows from the original table and re-insert them, and set pctfree properly, and the problem will be fixed now and into the future.

HJR Received on Tue Apr 27 2004 - 16:00:18 CDT