If we are going to change that portion of the code, we may as well go a bit forward and allow any expressions to be fetched from a foreign server (obviously, if that server is capable of doing so). It will help, when we come to aggregate push-down or whole query push-down (whenever that happens). So, instead of attr_needed, which restricts only the attributes to be fetched, why not to targetlist itself?
On Mon, Jun 30, 2014 at 7:18 PM, Tom Lane <t...@sss.pgh.pa.us> wrote: > Etsuro Fujita <fujita.ets...@lab.ntt.co.jp> writes: > > Done. I think this is because create_foreignscan_plan() makes reference > > to attr_needed, which isn't computed for inheritance children. > > I wonder whether it isn't time to change that. It was coded like that > originally only because calculating the values would've been a waste of > cycles at the time. But this is at least the third place where it'd be > useful to have attr_needed for child rels. > > regards, tom lane > -- Best Wishes, Ashutosh Bapat EnterpriseDB Corporation The Postgres Database Company