Re: Re-create dependent views on ALTER TABLE ALTER COLUMN ... TYPE?

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Stephen Frost <sfrost@snowman.net>, ash <ash@commandprompt.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2014-06-03T20:43:58Z
Lists: pgsql-hackers
Robert Haas <robertmhaas@gmail.com> writes:
> On Tue, Jun 3, 2014 at 10:14 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> I'm guessing you did not read
>> http://www.postgresql.org/message-id/18723.1401734537@sss.pgh.pa.us

> Argh, sorry, I saw that go by and it went past my eyes but obviously I
> didn't really absorb it.  I guess we could do it that way.  But it
> seems like quite a hassle to me; I think we're going to continue to
> get complaints here until this is Easy.  And if it can't be made Easy,
> then we're going to continue to get complaints forever.

Well, my vision of it is that it *is* easy, if you're using the tool
(or, perhaps, one of several tools), and you have a case that doesn't
really require careful semantic review.  But trying to build this sort
of thing into the backend is the wrong approach: it's going to lead
to unpleasant compromises and/or surprises.  And we'd still have to
build that tool someday.

			regards, tom lane


Commits

  1. Allow CREATE OR REPLACE VIEW to add columns to the _end_ of the view.