Message-ID: <de6c5d45-7ce9-42bd-9528-97fdf4ffa0e4@gmail.com>
Date: Mon, 9 Mar 2026 21:21:44 +0100
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Content-Language: en-US
To: "pgsql-general@lists.postgresql.org" <pgsql-general@lists.postgresql.org>
From: Pierre Forstmann <pierre.forstmann@gmail.com>
Subject: =?UTF-8?Q?LISTAGG_=C3=A0_la_Oracle_in_PostgreSQL?=
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <https://www.postgresql.org/message-id/de6c5d45-7ce9-42bd-9528-97fdf4ffa0e4%40gmail.com>
Precedence: bulk

Hello,

I can write a LISTAGG aggregate for:

create table emp(deptno numeric, ename text);

SELECT deptno, LISTAGG(ename, ','::text ORDER BY ename) AS employees 
FROM   emp GROUP BY deptno ORDER BY deptno;

I would like to know if is possible to create an aggregate LISTAGG that 
would work like in Oracle:

SELECT deptno,
        listagg(ename, ',') WITHIN GROUP (ORDER BY ename) AS employees
FROM emp
GROUP BY deptno
ORDER BY deptno;

I failed and IA also failed. Claude says:

It is not possible to exactly replicate listagg(ename, ',') WITHIN GROUP 
(ORDER BY ename) as a custom PostgreSQL aggregate
because PostgreSQL strictly forbids ungrouped columns as direct 
arguments to ordered-set aggregates.

Do you agree ?