> -----Original Message-----
> From: Andreas Pflug [mailto:pgadmin@pse-consulting.de]
> Sent: 20 September 2004 17:43
> To: Dave Page
> Cc: pgadmin-hackers@postgresql.org; Andreas Pflug
> Subject: Re: [pgadmin-hackers] CVS Commit by andreas: todo update
>
> Dave Page wrote:
> >
> >
> >
> >>-----Original Message-----
> >>From: pgadmin-hackers-owner@postgresql.org
> >>[mailto:pgadmin-hackers-owner@postgresql.org] On Behalf Of
> >>cvs@cvs.pgadmin.org
> >>Sent: 20 September 2004 16:25
> >>To: pgadmin-hackers@postgresql.org
> >>Subject: [pgadmin-hackers] CVS Commit by andreas: todo update
> >>
> >>+ <li>EditGrid: allow editing if unique key existent, but
> >
> > no PK/OID
> >
> > You realise that'll only work if there is a not null
> contraint as well?
>
> Not really.
Not really you didn't realise, or not really you don't agree? Unique
indexes can still contain multiple null values on the same column, hence
a not null constraint is required to make them useful as a row
identifier.
> Maybe we should accept PK for educational purposes only.
I'd rather not prevent the user from shooting his own foot in quite such
a nannying way, but on a more general theme, what about having a bunch
of tips in appropriate places to help the newbie? Each one could have
it's own 'don't show me this again' check box, along with a global
option to turn them off. In this case, when creating a table without a
pkey it could pop-up and warn about the problem and give the user the
chance to correct the problem.
Regards, Dave.