Horton

Re: Default Partition for Range

Robert Haas <robertmhaas@gmail.com>

View thread View on postgr.es
From: Robert Haas <robertmhaas@gmail.com>
To: Beena Emerson <memissemerson@gmail.com>
Cc: Jeevan Ladhe <jeevan.ladhe@enterprisedb.com>, Rahila Syed <rahilasyed90@gmail.com>, Dilip Kumar <dilipbalaut@gmail.com>, Amit Kapila <amit.kapila16@gmail.com>, Ashutosh Bapat <ashutosh.bapat@enterprisedb.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2017-08-04T14:18:45Z
Lists: pgsql-hackers 
On Mon, Jul 31, 2017 at 8:28 AM, Beena Emerson <memissemerson@gmail.com> wrote:
> Thanks for informing.
> PFA the updated patch.
> I have changed the numbering of enum PartitionRangeDatumKind since I
> have to include DEFAULT as well. If you have better ideas, let me
> know.

Why do we need to introduce PARTITION_RANGE_DATUM_DEFAULT at all?  It
seems to me that the handling of default range partitions ought to be
similar to the way a null-accepting list partition is handled -
namely, it wouldn't show up in the "datums" or "kind" array at all,
instead just showing up in PartitionBoundInfoData's default_index
field.

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

Commits

  1. Allow a partitioned table to have a default partition.

  2. Add new files to nls.mk and add translation markers

  3. Improve release note text about set-returning-function changes.