On Wed, May 25, 2011 at 1:37 PM, Hitoshi Harada <umi.tanuki@gmail.com> wrote:
>> How do you decide whether or not to push down?
>
> Yeah, that's the problem. In addition to the conditions of join-qual
> == grouping key && outer is unique on qual, we need some criteria if
> it should be done. At first I started to think I can compare cost of
> two different plan nodes, which are generated by calling
> subquery_planner() twice. But now my plan is to apply some heuristics
> like that join qual selectivity is less than 10% or so. I either don't
> like magic numbers but given Query restructuring instead of
> PlannerInfo (which means we cannot use Path) it is only left way. To
> get it work is my first goal anyway.
I think getting it working is probably a good first goal. I am not
really sure that we want to commit it that way, and I think my vote
would be for you to work on the approach we discussed before rather
than this one, but it's your project, and I think you'll probably
learn enough in getting it working that it will be a step forward in
any case. The planner is complex enough that it's worth trying to get
something that works, first, before trying to make it perfect.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company