Mark Harrison writes:
> Is there a way to optimize count(*) such that it does not have> to do a sequential scan?
No. If you need to count a lot, you need to store the informationseparately.
-- Peter Eisentraut peter_e(at)gmx(dot)net