From: Tom Lane <tgl@sss.pgh.pa.us>
To: Maxim Gasumyants <m@gasumyants.com>
cc: pgsql-bugs@lists.postgresql.org
Subject: Re: Generated column and partitioning bug
In-reply-to: <95418DF5-D567-4FFC-A77A-DCE9B985692F@gasumyants.com>
References: <95418DF5-D567-4FFC-A77A-DCE9B985692F@gasumyants.com>
Comments: In-reply-to Maxim Gasumyants <m@gasumyants.com>
	message dated "Fri, 04 Feb 2022 14:22:35 +0200"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <415296.1643997885.1@sss.pgh.pa.us>
Date: Fri, 04 Feb 2022 13:04:45 -0500
Message-ID: <415297.1643997885@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/415297.1643997885%40sss.pgh.pa.us>
Precedence: bulk

Maxim Gasumyants <m@gasumyants.com> writes:
> I am running latest release of PostgreSQL 12.

Are you certain it's the latest?  This looks suspiciously like
some Var-numbering bugs we fixed a year or so ago.

I failed to reproduce any problem in 12.9 on the basis of the
info you supplied.  I suppose that there might be some additional
conditions needed, like a dropped column in the parent table,
but experimentation didn't find it.

It might be useful to show the results of

select attname, attnum, attisdropped from pg_attribute
where attrelid = 'products_product_offers'::regclass and attnum > 0;

Also, if you can make a self-contained SQL script that exhibits
the problem for you, that would be very useful.

			regards, tom lane