Message-ID: <3E019829.204@klaster.net>
Date: Thu, 19 Dec 2002 10:58:01 +0100
From: Tomasz Myrta <jasiek@klaster.net>
User-Agent: Mozilla/5.0 (Windows; U; Win 9x 4.90; en-US;
	rv:1.2b) Gecko/20021016
MIME-Version: 1.0
To: Gary Stainburn <gary.stainburn@ringways.co.uk>
Cc: pgsql-sql@postgresql.org
Subject: Re: references table(multiple columns go here)
References: <200212181625.11316.gary.stainburn@ringways.co.uk>
	<200212181718.42870.gary.stainburn@ringways.co.uk>
	<8026.1040233829@sss.pgh.pa.us>
	<200212190928.19736.gary.stainburn@ringways.co.uk>
In-Reply-To: <200212181625.11316.gary.stainburn@ringways.co.uk>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Gary Stainburn wrote:

> That did the trick.  However, I now have another problem with the 
> constraint
> complaining about there not being an index to refer to.  However, 
> there is.
> Output below:
>
> create table ranks (
> rid     	int4 default nextval('ranks_rid_seq'::text) unique not null,
> rdid		character references depts(did), -- department
> rrank		int4 not null,		-- departmental rank
> rdesc		character varying(40)	-- Rank Description

,CONSTRAINT ranks_pkey PRIMARY KEY (rid,rrank)
or
,CONSTRAINT ranks_unq UNIQUE (rid,rrank)

>
> );

Create primary key on two fields in table ranks, or at least create 
unique constraint on them.

If rid is unique, why do you use two fields as foreign key? "rid" is 
enough. You can get rid of "rrank" in table jobtypes.

And one more question - why you don't use the same names in all tables?
"did" instead of "did" "rdid" "jdid" ? It's much easier to create joins 
when using the same names.
Tomasz Myrta