Re:

* Ed Tyrrill (tyrrill_ed(at)emc(dot)com) wrote:
> It seems to me that the first plan is the optimal one for this case, but
> when the planner has more information about the table it chooses not to
> use it. Do you think that if work_mem were higher it might choose the
> first plan again?

Seems likely to me. You understand that you can set the work_mem
whenever you want, right? It's a GUC, so you could issue a 'set
work_mem = blah' in the application code right before and right after
(assuming you're going to continue using the session) this particular
query, or just do it in a seperate session using 'explain' to play
around with what the planner does given different arguments.

'explain's pretty cheap/easy, and you can play around with various
settings to see what PG will do in various cases. Of course, you won't
know the runtimes without doing 'explain analyze', but I think you have
a good idea of the best plan for this query already...

Thanks,

Stephen