Archives of the TeradataForum
Message Posted: Thu, 22 Jan 2004 @ 21:59:34 GMT
<-- Anonymously Posted: Thursday, January 22, 2004 16:58 -->
Even I have encountered this. Statistics were collected on all the 7 tables joined together and all of them had Secondary Indexes. There were 4 Inner Joins and 2 Left Outer Joins in the SQL. And the explain always returned No CONFIDENCE. The tables were less than 2 GB in size.
I found that one of the tables was badly distributed (60% rows on 1 AMP) and all the spool files were forced that AMP. Stats somehow did not recognise the skewed distribution.
I then dropped the STATS which forced the skewed table to get redistributed and the skew got removed. We are on V2R5.0.
Looking for more info on this.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|