Archives of the TeradataForum

Message Posted: Fri, 14 Oct 2005 @ 11:00:59 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Glen and Ole,

The Statistics on all indexes are current, but the optimizer's row estimates are way off. We are on version V2R5.1.2. Ole, you are correct, our fact tables are the AJIs in this cases. The NUSIs I'm having trouble with exist on the AJIs. In addition, we are ordering our AJIs by a week value (instead of hash), and I am very suspicious about how the optimizer is costing this feature (which is always used in conjunction with the NUSI access.) In addition, we do use milestones, so, skewing combined with milestones could be a culprit.

I appreciate your input. I'll keep digging.

Thanks,

Clay

	<Prev	Next>		<<First	<Prev	Next>	Last>>