On Tue, 2006-05-02 at 00:27 -0500, Brandon Black wrote:
> I tried things like the above with small test data sets against cvs
> just now on my home machine, and constraint exclusion doesn't seem to
> apply here (even if all of the joined jstuff rows have ids which only
> match the constraint for basic_sub3, all basic_subX's seem to get
> scanned, as is the case I'm seeing in my real code against 8.1.3). Is
> this sort of dynamic constraint exclusion on the radar?
(You should use life size data for your testing.)
If you think the plan you get can be improved, post the EXPLAIN (maybe
EXPLAIN ANALYZE) and say how you think it can be improved. It's a lot
easier to consider concrete requirements.
-- Simon Riggs EnterpriseDB http://www.enterprisedb.com