On Fri, Feb 27, 2009 at 11:57 AM, Gregory Stark <stark@xxxxxxxxxxxxxxxx> wrote: > Uh, we get a lot of really mangled SQL and explain plans -- I don't see > anything wrong with these. If the question was unclear it sounds like it's > just because it's a fairly subtle problem and was hard to describe. Needing > two cracks at describing the problem is pretty much par for the course here. > > I haven't tested the query to see what's going on but if the problem is due to > random() then in 8.4 you could use WITH to guarantee that the subquery is > executed precisely once and the results reused as-is subsequently. In that case, actually I need to get a random array of packages that's different as much as possible for every account-id. I'll try to create simpler example, with some script that would generate data for you. -- GJ -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
