|
|
Archives of the TeradataForum
Message Posted: Mon, 24 Feb 2004 @ 04:35:06 GMT
Subj: | | Re: Multi-column stats for secondary index elimination |
|
From: | | Stubbs, Donald |
I wasn't in a situation where I was trying to save space by using multi-column statistics, but multi-column stats will sometimes make big
differences in plans the optimizer chooses. In queries where you have qualifications on multiple columns, where before all the optimizer had were
stats on individual columns, it can now have stats on the combination of columns (values), and it can estimate the spool sizes much better. I've
seen some dramatic improvements from this feature (not in all cases, mind you). And if the optimizer indicates it's not using the index, then you
can drop it and save the space.
Don Stubbs
Teradata Advanced Certified Professional
| |