(2018/09/18 23:20), Jinhua Luo wrote:
> Sorry, the example is not so proper. I just think even if it's a
> simple example, e.g. join two co-located tables, the planner should
> work out to push down it. Can you confirm the postgresql could detect
> co-located tables on the same foreign server and push down queries on
> them? Could you give an actual example or point out the relevant
> source code paths for reference?
>
> (Let me clarify the context of this question, if the planner supports
> co-located push-down, then it's meaningful for manual sharding via
> partitioning to remote tables, where it's mostly necessary to join two
> or more co-located parent tables in complex queries. If not, the
> postgresql instance on which the parent tables are placed (let's say
> it's a coordinator node) would be likely the bottleneck.)
You might want to check partitionwise join functionality as well, which
we have in PG11 [1].
Best regards,
Etsuro Fujita
[1]
https://git.postgresql.org/gitweb/?p=postgresql.git;a=commit;h=f49842d1ee31b976c681322f76025d7732e860f3