Renaud Tthonnart <thonnart(at)amwdb(dot)u-strasbg(dot)fr> writes:
> Nested Loop (cost=0.00..56.67 rows=3 width=20)
> -> Seq Scan on observation o (cost=0.00..30.00 rows=1 width=4)
> -> Seq Scan on exposure_epic e (cost=0.00..22.50 rows=333 width=16)
Do a VACUUM ANALYZE, for starters. These results look like the planner
is working with the initial default estimates for a never-vacuumed table.
> I haven't create any index
Indexes might be a good idea too.
http://www.postgresql.org/devel-corner/docs/postgres/indices.html
has a good intro to the basics.
regards, tom lane