Date: Thu, 4 Jan 2018 19:01:33 -0300
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
To: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>
Cc: Robert Haas <robertmhaas@gmail.com>,
	Amit Langote <Langote_Amit_f8@lab.ntt.co.jp>,
	Jesper Pedersen <jesper.pedersen@redhat.com>,
	David Rowley <david.rowley@2ndquadrant.com>,
	Simon Riggs <simon@2ndquadrant.com>,
	Pg Hackers <pgsql-hackers@postgresql.org>,
	Maksim Milyutin <milyutinma@gmail.com>
Subject: Re: [HACKERS] Proposal: Local indexes for partitioned table
Message-ID: <20180104220133.6sa4zpf5liurxhte@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <584730ad-6736-10bc-b5fa-251d4873b8ad@2ndquadrant.com>
User-Agent: NeoMutt/20170306-137-4415bd-dirty (1.8.0)
Precedence: bulk

Peter Eisentraut wrote:
> On 1/4/18 12:00, Robert Haas wrote:
> >> The catalog representations of partitioned tables and partitioned
> >> indexes are completely different, which may or may not be desirable.
> > 
> > How so?
> 
> If someone wants to write a query, show me all the partitions of this
> table versus show me all the partitions of this index, intuitively,
> those could be the same, different only by some relkind references.

Well, this "indparentidx" stuff I came up with is a little bit strange.
Perhaps we could use pg_inherits instead, but I think the general view
is that pg_inherits is on its way out.

Tangentially: I didn't like very much that I added a new index to
pg_index to support this feature.  I thought maybe it'd be better to
change the index on indrelid to be on (indrelid,indparentidx) instead,
but that doesn't seem great either because it bloats that index which is
used to support common relcache operations ...


(The more I think of this, the more I believe that pg_inherits is a
better answer.  Opinions?)

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services