Date: Mon, 3 Nov 2025 18:22:11 +0100
From: hubert depesz lubaczewski <depesz@depesz.com>
Sender: depesz@depesz.com
To: dfgpostgres <dfgpostgres3@gmail.com>
Cc: pgsql-general@lists.postgresql.org
Subject: Re: PG Unpivot ?
Message-ID: <aQjkwyDjsz8Cgp0X@depesz.com>
Reply-To: depesz@depesz.com
References: <CAAcmDX_KU4A1DWzKmO4XacVQ72+0KZ9rZSvQwHkCWgpqB-Lb2w@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <CAAcmDX_KU4A1DWzKmO4XacVQ72+0KZ9rZSvQwHkCWgpqB-Lb2w@mail.gmail.com>
Archived-At: <https://www.postgresql.org/message-id/aQjkwyDjsz8Cgp0X%40depesz.com>
Precedence: bulk

On Mon, Nov 03, 2025 at 12:18:55PM -0500, dfgpostgres wrote:
> psql (13.2, server 15.3) on linux
> 
> I think they call this "unpivot" in MSSQL ?
> 
> How can I get an sql query to return  one line per column with... an ID,
> column name and value.  the ctid for the id field is fine.
> 
> Example:
> dvdb=# create table unpivot (intcol integer, floatcol float, strcol
> varchar);
> CREATE TABLE
> dvdb=# insert into unpivot (intcol,floatcol,strcol) values
> (1,1.1,'one'),(2,2.2,'two'),(3,3.3,'three');
> INSERT 0 3
> dvdb=# select * from unpivot;
>  intcol | floatcol | strcol
> --------+----------+--------
>       1 |      1.1 | one
>       2 |      2.2 | two
>       3 |      3.3 | three
> (3 rows)
> I want 9 records returned, each row with 3 cols, 1st col is the ctid,
> second is the column name, third is the val.

I think it should work:

select u.ctid, e.* from unpivot u, to_jsonb(u) j, jsonb_each_text(j) e;

Best regards,

depesz