From: Tom Lane <tgl@sss.pgh.pa.us>
To: Pavel Stehule <pavel.stehule@gmail.com>
cc: james.inform@pharmapp.de,
        PostgreSQL mailing lists <pgsql-bugs@lists.postgresql.org>
Subject: Re: BUG #17387: Working in PG13 but not in PGH14: array_agg(RECORD)
In-reply-to: <2807139.1643406198@sss.pgh.pa.us>
References: <17387-5dfe54b988444963@postgresql.org>
 <CAFj8pRCdDOs9w-u7x2LHoN-uUVOQqj4SH+VfBWA9BCHBDpckgw@mail.gmail.com>
 <CAFj8pRDMMTXX1B56P=7cvJvDSOYuhrVE_PBch7EFOV8-qOHNzQ@mail.gmail.com>
 <2807139.1643406198@sss.pgh.pa.us>
Comments: In-reply-to Tom Lane <tgl@sss.pgh.pa.us>
	message dated "Fri, 28 Jan 2022 16:43:18 -0500"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2808393.1643406921.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 28 Jan 2022 16:55:21 -0500
Message-ID: <2808394.1643406921@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/2808394.1643406921%40sss.pgh.pa.us>
Precedence: bulk

I wrote:
> ... which was non-ambiguous because in this usage, anyelement
> wouldn't match an array type.  I wonder why that's not
> happening with the anycompatible family?

Poking further, this case still works:

regression=3D# select array[1] || array[2];
 ?column? =

----------
 {1,2}
(1 row)

so we didn't break it completely (I rather imagine we have
regression tests that would have noticed that).  Also,
you can still concatenate arrays of known composite types:

regression=3D# select array_agg(t) || array_agg(t) from int8_tbl t;
                                                                          =
      =

                                                        ?column?          =
      =

                                                                          =
      =

                                         =

--------------------------------------------------------------------------=
------
--------------------------------------------------------------------------=
------
--------------------------------------------------------------------------=
------
-----------------------------------------
 {"(123,456)","(123,4567890123456789)","(4567890123456789,123)","(45678901=
234567
89,4567890123456789)","(4567890123456789,-4567890123456789)","(123,456)","=
(123,4
567890123456789)","(4567890123456789,123)","(4567890123456789,456789012345=
6789)"
,"(4567890123456789,-4567890123456789)"}
(1 row)

So it seems like this is specific to type record[] somehow.
Odd.

			regards, tom lane