Ceri Storey <cez-misc.pgsql-perf@necrofish.org.uk> writes:
> Although, as I've just found, another bottleneck is the title table.
> PostgreSQL seems to inst on doing a Seq Scan on the entire table.
> -> Seq Scan on tid (cost=0.00..20.00 rows=1000 width=8) (actual time=0.028..10.457 rows=17 loops=1)
It doesn't look like you've ever vacuumed or analyzed "tid" --- those
are the default cost and rows estimates. Although I'm unsure whether
the plan would change much if you had.
regards, tom lane