Date: Tue, 10 Sep 2024 08:35:48 -0700 (PDT)
From: Rich Shepard <rshepard@appl-ecosys.com>
To: pgsql-general@postgresql.org
Subject: Re: Removing duplicate rows in table
In-Reply-To: <5872cd30-9a2b-4b4c-82c4-48228197e872@a-kretschmer.de>
Message-ID: <265771f-f2c8-3c98-aebf-6239da97f3b2@appl-ecosys.com>
References: <e4688faf-de51-93ef-2db-f2acacbd658@appl-ecosys.com> <5872cd30-9a2b-4b4c-82c4-48228197e872@a-kretschmer.de>
MIME-Version: 1.0
Content-Type: text/plain; format=flowed; charset=ISO-8859-15
Content-Transfer-Encoding: 8BIT
Archived-At: <https://www.postgresql.org/message-id/265771f-f2c8-3c98-aebf-6239da97f3b2%40appl-ecosys.com>
Precedence: bulk

On Tue, 10 Sep 2024, Andreas Kretschmer wrote:

> you can use the hidden ctid-column:
>
> postgres=# create table demo (id int, val text);
> CREATE TABLE
> postgres=# insert into demo values (1, 'test1');
> INSERT 0 1
> postgres=# insert into demo values (1, 'test1');
> INSERT 0 1
> postgres=# insert into demo values (1, 'test1');
> INSERT 0 1
> postgres=# insert into demo values (1, 'test1');
> INSERT 0 1
> postgres=# select ctid, * from demo;
>  ctid  | id |  val
> -------+----+-------
>  (0,1) |  1 | test1
>  (0,2) |  1 | test1
>  (0,3) |  1 | test1
>  (0,4) |  1 | test1
> (4 rows)
>
> postgres=# with my_ctid as (select min(ctid) from demo where id = 1 and val = 
> 'test1') delete from demo using my_ctid where id=1 and val='test1' and ctid 
> != my_ctid.min;
> DELETE 3
> postgres=# select ctid, * from demo;
>  ctid  | id |  val
> -------+----+-------
>  (0,1) |  1 | test1
> (1 row)

Thanks, Andreas.

Rich