Re: problems with foreign keys on partitioned tables

On Sat, Jan 19, 2019 at 7:16 AM Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> wrote:
> Thanks, this is better. There were a few other things I didn't like, so
> I updated it. Mostly, two things:
>
> 1. I didn't like a seqscan on pg_trigger, so I turned that into an
> indexed scan on the constraint OID, and then the other two conditions
> are checked in the returned tuples. Also, what's the point on
> duplicating code and checking how many you deleted? Just delete them
> all.

Yeah, I didn't quite like what that code looked like, but it didn't
occur to me that there's an index on tgconstraint.

It looks much better now.

> 2. I didn't like the ABI break, and it wasn't necessary: you can just
> call createForeignKeyActionTriggers directly. That's much simpler.

OK.

> I also added tests. While running them, I noticed that my previous
> commit was broken in terms of relcache invalidation. I don't really
> know if this is a new problem with that commit, or an existing one. The
> fix is 0001.

Looks good.

Thanks,
Amit