Horton

Re: No-rewrite timestamp<->timestamptz conversions

Noah Misch <noah@leadboat.com>

View thread View on postgr.es
From: Noah Misch <noah@leadboat.com>
To: Simon Riggs <simon@2ndquadrant.com>
Cc: Josh Berkus <josh@agliodbs.com>, Andres Freund <andres@2ndquadrant.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2019-02-26T20:12:35Z
Lists: pgsql-hackers 
On Tue, Feb 26, 2019 at 02:29:01PM +0000, Simon Riggs wrote:
> Looks good, would need docs.

The ALTER TABLE page just says "old type is either binary coercible to the new
type or an unconstrained domain over the new type."  Avoiding rewrites by way
of a prosupport function or the at-timestamp-v2.patch approach is essentially
another way to achieve binary coercion.  So far, we haven't documented the
individual data types affected.  Since we don't mention e.g. varchar(x) ->
varchar(x+k) explicitly, I plan not to mention timestamp explicitly.

Commits

  1. Avoid some table rewrites for ALTER TABLE .. SET DATA TYPE timestamp.

  2. Disable transforms that replaced AT TIME ZONE with RelabelType.

  3. Add transform functions for AT TIME ZONE.