Content-Type: text/plain;
  charset="iso-8859-1"
From: Josh Berkus <josh@agliodbs.com>
Reply-To: josh@agliodbs.com
Organization: Aglio Database Solutions
To: Robert Treat <xzilla@users.sourceforge.net>, elein@varlena.com,
	elein <elein@sbcglobal.net>
Subject: Re: [SQL] rewriting values with before trigger
Date: Fri, 25 Apr 2003 10:20:48 -0700
User-Agent: KMail/1.4.3
Cc: "pgsql-general@postgresql.org" <pgsql-general@postgresql.org>,
	pgsql-sql@postgresql.org
References: <1051201544.4753.522.camel@camel>
	<200304250033.h3P0XuDl088102@pimout1-ext.prodigy.net>
	<200304250003.24506.xzilla@users.sourceforge.net>
In-Reply-To: <200304250003.24506.xzilla@users.sourceforge.net>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Message-Id: <200304251020.48678.josh@agliodbs.com>

Robert,

> You misunderstood. I don't think it's a bug in postgresql,  it's a bug in=
=20
the=20
> application that is hitting against my database. When it doesn't have a=
=20
value=20
> for the timestamp field, it either needs to drop it from the insert statm=
ent=20
> or convert it to null; not send a ''

Incidentally, this sort of problem is why most of my apps are based on "pus=
h"=20
data functions.    i.e., instead of the client calling:
INSERT INTO foo VALUES ( id, bar1, bar2, bar3 );
it calls
SELECT df_modify_foo ( id, bar1, bar2, bar3 );

Data-push functions allow me to do a whole array of validation and custom=
=20
error message return that would be impractical with triggers.   It also=20
allows me to build security checks in to the back-end, via:

SELECT df_modify_foo ( user_id, session_key, id, bar1, bar2, bar3 )

... allowing me to check all of the following things:
Does the user have a valid session?
Does the user have rights to foo?
Does the user have a lock on foo?
Is this a new foo record, or a modified one?
etc.

--=20
-Josh Berkus
 Aglio Database Solutions
 San Francisco