On Tue, 10 Jan 2006, Andrea Arcangeli wrote:
> I see. I can certainly fix it by stopping using LIKE. But IMHO this
> remains a bug, since until the statistics about the numberof matching
> rows isn't estimated well, you should not make assumptions on LIKE/NOT
> LIKE. I think you can change the code in a way that I won't have to
> touch anything, and this will lead to fewer surprises in the future IMHO.
I doubt it, since I would expect that this would be as large a
pessimization for a larger fraction of people than it is an optimization
for you.
