Re: Fix for pg_upgrade and invalid indexes - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: Fix for pg_upgrade and invalid indexes
Date
Msg-id 20130330033513.GB12134@momjian.us
Whole thread Raw
In response to Re: Fix for pg_upgrade and invalid indexes  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Fix for pg_upgrade and invalid indexes  (Bruce Momjian <bruce@momjian.us>)
List pgsql-hackers
On Fri, Mar 29, 2013 at 07:03:05PM -0400, Tom Lane wrote:
> Andres Freund <andres@2ndquadrant.com> writes:
> > Those columns cannot be NULL, so using IS DISTINCT FROM seems a bit
> > clumsy.
>
> That was what I started to write, too, but actually I think the IS
> DISTINCT is correct and the RIGHT JOIN should be a LEFT JOIN.  Note
> that the query appears to be intended to collect regular tables as
> well as indexes.  (As patched, that's totally broken, so I infer
> Bruce hasn't tested it yet.)

Yes, I only ran my simple tests so far --- I wanted to at least get some
eyes on it.  I was wondering if we ever need to use parentheses for
queries that mix normal and outer joins?  I am unclear on that.

Attached is a fixed patch that uses LEFT JOIN.  I went back and looked
at the patch that added this test and I think the patch is now complete.
I would like to apply it tomorrow/Saturday so it will be ready for
Monday's packaging, and get some buildfarm time on it.

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

  + It's impossible for everything to be true. +

Attachment

pgsql-hackers by date:

Previous
From: Amit Kapila
Date:
Subject: Re: Getting to 9.3 beta
Next
From: Alexander Korotkov
Date:
Subject: Re: Getting to 9.3 beta