 |
 |
Archives of the TeradataForum
Message Posted: Mon, 11 Oct 2004 @ 14:59:26 GMT
| Subj: | | Re: Hey, I've got a question: Sampling Statistics |
| |
| From: | | David Clough |
Yonina Schuchman writes:
| | We have experienced the same problems with sampling. We noted really bad access paths, so we basically gave up on it for now. | |
Yonina, did you only apply sampling to unique or nearly unique columns, or did you apply it to low-cardinality columns as well?
I ask because, as I understand it, it only provides benefit to nearly unique values, and it was something I was about to employ myself - now
I'm not so sure!
Dave Clough
Database Designer
Database Design Group
| |
