From: Aarni =?iso-8859-1?q?Ruuhim=E4ki?= <aarni@kymi.com>
Reply-To: aarni@kymi.com
Organization: KYMI.com
To: pgsql-sql@postgresql.org
Subject: Re: Seeking quick way to clone a row, but give it a new pk.
Date: Thu, 8 Feb 2007 11:32:58 +0200
User-Agent: KMail/1.6.2
References: <45CA50E6.6090006@obviously.com>
	<20070207181622.I3035@bravo.pjkh.com>
	<45CACF12.1060800@obviously.com>
In-Reply-To: <45CACF12.1060800@obviously.com>
Cc: Philip Hallstrom <postgresql@philip.pjkh.com>
MIME-Version: 1.0
Content-Disposition: inline
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Message-Id: <200702081132.58609.aarni@kymi.com>

On Thursday 08 February 2007 09:19, Bryce Nesbitt wrote:
> >
> > INSERT INTO mytable SELECT * FROM mytable WHERE pk =3D 123;
> >
> > Or something close to that... I suspect if you changed the '*' to the
> > columns you wanted you could also work in the other columns you want
> > to change as well...
>
> But that will violate the unique primary key constraint:
>
> insert into xx_plan_rule select * from xx_plan_rule where rule_id=3D9;
> ERROR:  duplicate key violates unique constraint "xx_plan_rule_pkey"

It will, because you are copying all columns, including the pk.

Try:

INSERT INTO mytable (colname_1, colname_2, colname_3)=20
SELECT (colname_1, colname_2, colname_3)=20
=46ROM mytable WHERE pk =3D 123;

BR,
=2D-=20
Aarni Ruuhim=E4ki