Re: pg_dump emits ALTER TABLE ONLY partitioned_table

On Tue, Apr 25, 2017 at 12:26 PM, Stephen Frost <sfrost(at)snowman(dot)net> wrote:
> Interesting. Seems like the question is really what we mean by "ONLY"
> here. For my 2c, at least, if we can check that all of the partitions
> already have the constraint enforced, such that the only thing we're
> changing is the partitioned table, then that's exactly what "ONLY"
> should do. That would require a bit of rework, presumably, of how that
> keyword is handled but the basics are all there.

Well, I'm not saying that couldn't be done, but it doesn't add any
capability that we don't have today, so from my point of view it seems
fairly pointless. However, we can leave that for another day; for
now, I think we should focus on fixing the pg_dump bugs that this
thread is about.

> I'm planning to push just this patch to make the backend accept the
> ALTER TABLE ONLY when no partitions exist later this afternoon, ...

Cool.

> but the
> work here isn't done as noted in my other email.

Right. I'd just like to get the work done as soon as we reasonably
can. Tempus fugit, and all that.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company