Archives of the TeradataForum
Message Posted: Thu, 04 Aug 2005 @ 18:58:10 GMT
Subj: | | Re: Optimizer and Transitive Property of Equality |
|
From: | | Hassinger, Bill |
Although I have not made a study, from experience I have noticed that a centric join order appears to performs more rapidly than a linear join
order.
For example
Where a=b
And a=c
And a=d
And a=e
Is faster than
Where a=b
And b=c
And c=d
And d=e
Below you indicated adding a=c and time was reduced. What if you removed b=c and only had a=b and a=c? Is the NUSI used then?
|