Date: Mon, 16 Mar 2026 15:15:14 -0500
From: Nathan Bossart <nathandbossart@gmail.com>
To: Corey Huinker <corey.huinker@gmail.com>
Cc: Sami Imseih <samimseih@gmail.com>, pgsql-hackers@lists.postgresql.org,
	michael@paquier.xyz
Subject: Re: Add starelid, attnum to pg_stats and leverage this in pg_dump
Message-ID: <abhk0oLSA7lQ62jJ@nathan>
References: <abBIu0lYyFYAq5fx@nathan>
 <CADkLM=d-=_hbOUsMc4u_ScLmSNox6EbzFCmaMn7MTh6f_SLU_A@mail.gmail.com>
 <CADkLM=d-yUR8JsM-UmCfeRAhVj+X9MYi0kz9=zQO4EoBkd8H8A@mail.gmail.com>
 <CAA5RZ0tiL01rYm7=WaDGLUYjvz+fQRqmaXyRUv9jKRnR3_6xDQ@mail.gmail.com>
 <CADkLM=fn=RmBD0-h+yjTxhpLtbK9s_QdvT-HsGtUj+-1N=6s0g@mail.gmail.com>
 <CAA5RZ0sf7vYav+Q5GwxGikBZSYUm+4haA5bHC0N67ct570HLcg@mail.gmail.com>
 <abRugSlNTmW0-9Um@nathan>
 <CADkLM=cfpFMpgOGPEyr7U+-Cxzvukap-_s8p-ijMTyRphP=oeg@mail.gmail.com>
 <abhXqx-xffsknv3B@nathan>
 <CADkLM=cuwGZU0i51vn+SUqptXB8QT8rhRRYshns+pX0KxczmAQ@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: 
 <CADkLM=cuwGZU0i51vn+SUqptXB8QT8rhRRYshns+pX0KxczmAQ@mail.gmail.com>
Archived-At: <https://www.postgresql.org/message-id/abhk0oLSA7lQ62jJ%40nathan>
Precedence: bulk

On Mon, Mar 16, 2026 at 04:04:23PM -0400, Corey Huinker wrote:
>> I left the expr_attnum stuff out.  It seems to make this patch quite large
>> and complicated, we don't plan to use it for the pg_dump patch, and I'm not
>> sure about showing users a "synthetic attnum" that seems to have no other
>> point of reference.  Would this information be useful in pg_dump somewhere?
>> I'm curious to hear more about the intent.
> 
> expr_attnum was something that Michael Paquier had lamented that the view
> didn't have. There is obviously no present need for it, as pg_dump isn't
> being modified for extended stats at all.

Okay.  I think I'll continue to leave this one out for now.

>> I didn't see much value in adding attnum here given the size of the changes
>> to the expected output it produces.
> 
> Same reasons for putting that in - people had lamented that we couldn't
> order the dump by attnum, and ordering by attname feels weird somehow.
> Again, we don't presently need it.

This note was about adding attnum to the pg_stats_stable view in the test.
I don't have any problem with adding it to pg_stats.

-- 
nathan