Joe Conway <mail@joeconway.com> writes:
> It's hard to tell from what you did post, but if the query returns a
> significant portion of the table then a seq scan is faster, and is
> properly picked by the optimizer. Please post EXPLAIN ANALYZE results.
And do a VACUUM ANALYZE beforehand. The explain estimates look like
the planner thinks all these tables are tiny.
> (and actually, this thread probably should be on the SQL or the PERFORM
> lists, not this one)
Right.
regards, tom lane