On Thu, Mar 14, 2019 at 3:36 PM Alvaro Herrera <alvhe...@2ndquadrant.com> wrote: > Well, I suppose that can be implemented as an optimization on top of > what we have, but I think that we should first get this feature right, > and later we can see about improving it.
Sure, not arguing with that. > In any case, since the RI > queries are run via SPI, any unnecessary partitions should get pruned by > partition pruning based on each partition's constraint. So I'm not so > certain that this is a problem worth spending much time/effort/risk of > bugs on. I doubt that partition pruning would just automatically DTRT in a case like this, but if I'm wrong, great! -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
