Subject: Re: SEQUENCE and PRIMARY KEY
From: Rod Taylor <rbt@rbt.ca>
To: Dani Oderbolz <oderbolz@ecologic.de>
Cc: pgsql-sql@postgresql.org
In-Reply-To: <3F03E3B6.80807@ecologic.de>
References: <200307021730.23659.ralf.werny@gmx.de>
	<1057160585.90653.88.camel@jester>  <3F03E3B6.80807@ecologic.de>
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature";
	boundary="=-jW7IPdsIGMc46LpWG0NL"
Message-Id: <1057234937.30368.26.camel@jester>
Mime-Version: 1.0
Date: 03 Jul 2003 08:22:17 -0400

--=-jW7IPdsIGMc46LpWG0NL
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

> >A trigger (maybe a rule) is the way to go about this.
> >
> >The alternative is to teach the bad clients to use DEFAULT rather than
> >NULL when they expect the GENERATOR to create the value for them.
> >
> Hmm, Rod, there really seems a demand for the GENERATOR feature :-)
> Do you see any performance problems with a trigger instead of a default?
> I cant imagine that its so much of a difference (ok, depending on the=20
> trigger, a
> lookup in the catalog is needed, but still it should be fast enough, righ=
t?)

A trigger written in C will have similar speeds as processing the
default.  Plpgsql isn't that slow, but it is slower -- probably not
noticeably for this operation.

--=-jW7IPdsIGMc46LpWG0NL
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (FreeBSD)

iD8DBQA/BB/46DETLow6vwwRAhhRAJ0bZuXSdowp7OMdflJWXaSbGhfnRQCeNXgn
8Lv85S1ztafZsZFVXn3L2z8=
=EzDu
-----END PGP SIGNATURE-----

--=-jW7IPdsIGMc46LpWG0NL--