Archives of the TeradataForum
Message Posted: Wed, 10 Jan 2001 @ 22:04:50 GMT
Now I'm not so sure either. I believe that the statement made in the book is true for the simple sample given, but I'm not so sure that it would be true in all cases. When I answered, I had in mind a more complex query with joins and a HAVING clause. Depending on what it might be trying to accomplish (and keep in mind that it's a contrived and vague example), wouldn't the majority of the work still have to be done before the SAMPLE would be applied?
So the bottom-line is probably that the only way to be sure is by checking the EXPLAIN - on a case-by-case basis.
I tried to find some kind of write-up about the ALL-AMP SAMPLE step without success. Does anybody have an idea if it's described anywhere? There used to be a paper called 'Explaining the EXPLAIN'. Does anybody know if it's been kept up-to-date and if it's available somewhere?
I can't answer why the optimizer gives a time estimate some of the time and not others - it probably has to do with the certainty that the optimizer has for its estimates.
Does anybody know if the time estimates are more accurate in V2R4? I remember that there was supposed to be an effort in V2R2 to give the time estimates have some kind of validity - but I think that the estimates were still pretty wild and couldn't be depended upon. Prior to that, the time estimates where pure nonsense.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|