On Fri, 9 Nov 2018 at 06:58, Robert Haas <robertmhaas@gmail.com> wrote:
> On Wed, Nov 7, 2018 at 9:43 PM Michael Paquier <michael@paquier.xyz> wrote:
> > On Thu, Nov 08, 2018 at 12:50:40PM +1300, David Rowley wrote:
> > > How about we record the tablespace option for the partitioned table in
> > > reltablespace instead of saving it as 0. Newly created partitions
> > > which don't have a TABLESPACE mentioned in the CREATE TABLE command
> > > should be created in their direct parent partitioned tables
> > > tablespace.
> >
> > I have seen enough complains on the mailing lists regarding the way
> > tablespaces are handled for partitioned tables and their partitions that
> > it looks like a very good idea to make the tablespace being inherited
> > automatically, by setting up reltablespace to a non-zero value even if
> > a partitioned table has no physical presence. Of course not on v11 or
> > older releases, just on HEAD. It is no good to have partitioned indexes
> > and partitioned tables being handling inconsistently for such things.
>
> +1.
Here's a patch for that. Parking here until January's commitfest.
--
David Rowley http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services