Subject: Re: ENUM like data type
From: Rod Taylor <pg@rbt.ca>
To: =?ISO-8859-1?Q?Mart=EDn_Marqu=E9s?= <martin@bugs.unl.edu.ar>
Cc: pgsql-sql@postgresql.org
In-Reply-To: <200506291021.10986.martin@bugs.unl.edu.ar>
References: <ED4E30DD9C43D5118DFB00508BBBA76EB167AF@neptun.sonorys.at>
	<200506291021.10986.martin@bugs.unl.edu.ar>
Content-Type: text/plain; charset=ISO-8859-1
Date: Wed, 29 Jun 2005 09:28:49 -0400
Message-Id: <1120051729.39304.113.camel@home>
Mime-Version: 1.0
Content-Transfer-Encoding: 8bit

On Wed, 2005-06-29 at 10:21 -0300, Martín Marqués wrote:
> El Mié 29 Jun 2005 09:40, KÖPFERL Robert escribió:
> > 
> > |
> > |I personally think that the ENUM data type is for databases 
> > |that are not well 
> > |designed. So, if you see the need for ENUM, that means you 
> > |need to re-think 
> > |your data design.
> > |
> > 
> > I disagree. In several relations (views of the world) one needs to have a
> > hand full of well defined values while
> > integers or bools are not appropriate and strings are too free form.
> > For example male female or true and false. Whilst the second has a well
> > known type, other things like a telephone number type don't have such type.
> > So a new one should be created instead of (varchar/ xy between 4 and 8). One
> > might argue that new 1:n relations with integrity shall be established but
> > this is in my eyes overkill for a constant number of values.
> > Also think of schema export without data. These values are part of the
> > schema data but not the schema itself which in this sense is wrong
> 
> Please, most of this can be done with CHECK(...).

Indeed. A CHECK constraint on a DOMAIN is an ENUM plus some.

--