Mike Winter <mike(dot)winter(at)frontlogic(dot)com> writes:
> MySQL does queries of this type orders of magnitudes faster than Postgres
> on large value lists, although I have no specific algorithmic solutions to
> offer for how to make it faster.
How large is "large", and what plan type are you getting (seq scan or
multiple index scan)? Is it possible that the cost comes from planner
overhead and not execution? Checking EXPLAIN ANALYZE reported time
against actual elapsed time (cf psql's \timing option) would tell.
regards, tom lane