Re: PostgreSQL 7.3: help on new CREATE TYPE - Mailing list pgsql-hackers

From elein
Subject Re: PostgreSQL 7.3: help on new CREATE TYPE
Date
Msg-id 200209240210.g8O2AZI5620942@pimout2-ext.prodigy.net
Whole thread Raw
In response to Re: [GENERAL] PostgreSQL 7.3: help on new CREATE TYPE  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Illustra did a very nice job with "composite types" which
correspond to these record types.  The composite types
were able to be used as a column type as jerome describes.
The subcolumns were accessed with dots.  This gave usschema.table.column.attribute
where of course attribute could itself be a composite type....
Well, ok, it had some drawbacks, too.

If we ever are serious about implementing this I would help
with discussing and/or writing the specs.  I can put together a nice spec.
When I get a break on my book project, I might just write it up anyway.

elein
elein@norcov.com

PS: Everyone please forgive me for reading list mail late and out of order...
I am in awe of anyone keeping up.

On Tuesday 17 September 2002 07:22, Tom Lane wrote:
> "Jerome Chochon" <jerome.chochon@ensma.fr> writes:
> > Can I use this new type like other user-type ?
> > CREATE TABLE person (his_name VARCHAR, his_adress adress);
> > ...where adress is CREATE TYPE adress AS (number int, street text,
> > country VARCHAR);
>
> Not at the moment, though that might be an interesting direction to
> pursue in future releases.  At present, the only thing such a type is
> useful for is to define the argument or result type of a function that
> takes or returns records.
>
>             regards, tom lane
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org


pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: BETA2 HOLD: was Re: NUMERIC's transcendental functions
Next
From: Gavin Sherry
Date:
Subject: Web site