Re: SQL:2011 PERIODS vs Postgres Ranges?

On Tue, Aug 6, 2019 at 11:07 AM Ibrar Ahmed <ibrar(dot)ahmad(at)gmail(dot)com> wrote:
> Thanks, Paul for the explanation. I think its good start, now I am looking at the
> range_agg patch to integrate that with that and test that.

Since we've started another commitfest, here is an updated version of
this patch. I've rebased it on the latest multirange patch (which is
rebased on the latest master). I've incorporated your feedback. I've
also added some progress on adding FOR PORTION OF to UPDATE and DELETE
(mostly UPDATE). That is even more WIP than the PK/FK work, because
I'm still working on the executor phase, so feel free to ignore it or
offer feedback. I've put the DML work in a separate patch file. (I'm
planning to roughly follow ON CONFLICT DO UPDATE for the ModifyTable
changes, since temporal DML is also an extra clause that transforms
your commands into something else. For example a temporal UPDATE could
become an UPDATE plus two INSERTs. I'm probably going to need some
help eventually getting the concurrency stuff right here though.) (I
think temporal DML makes sense to include in this patch because it is
required for cascading FKs. I think once everything is working I'll
give you a patch series that goes PKs - DML - FKs. Or even better
PERIODs - PKs - DML - FKs.)

Thanks!
Paul