"Kevin Grittner" <[EMAIL PROTECTED]> writes:
> Dang. Obviously, that's inverted. Also, I'd need to
> factor in the setup time. Bother. Are you sure we can't
> just make sure the test scripts operate against tables
> with accurate statistics?
Well, the point of my comment was that you should record the rowcounts
so that you know whether or not a particular test case was skewed by
inaccurate stats. Whether you choose to disregard those cases or try
to include them in the analysis is up to you --- but assuming that every
number submitted to you is unaffected by bad stats is just asking for
trouble.
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
