Archives of the TeradataForum
Message Posted: Wed, 18 Aug 2010 @ 16:59:54 GMT
Subj: | | Re: How to Overcome product join |
|
From: | | John_Wight |
Yeah - the TD optimizer is pretty awesome most of the time and gets better with each release. I remember in TD V1 (the really old TD) where
stats didn't really make that much difference in a lot of cases but since V2 and going forward they have become VERY important. There is a good
Orange Book by Carrie Ballinger on stats that you should all review. Also there are some good articles in the Teradata Developers Exchange if you
get time to sign up and use.
However, there are still some things that make you wonder like DISTINCT -vs- GROUP BY and the way it works. I did see something that in TD
13.10 where it will 'sort that out' and you won't have to 'study the data' to see which one to use - don't know the details though! I remember
Todd Walters saying some time ago that this was something to be addressed in a future release, so it seems like it finally bubbled up and got put
in. Looking forward to seeing the explain plans on each. If anyone has the pre-release of 13.10 (don't think it's GS until Sept/Oct??) it would
be interesting if you could try both, explain them and share that with us out of curiosity!
But then - THAT killed the cat;-)
JK
|