From: | Rafal Pietrak <rafal(at)zorro(dot)isa-geek(dot)com> |
---|---|
To: | SunWuKung <Balazs(dot)Klein(at)t-online(dot)hu> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: A VIEW mimicing a TABLE |
Date: | 2006-12-14 11:52:23 |
Message-ID: | 1166097143.27564.114.camel@zorro.isa-geek.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Thu, 2006-12-14 at 02:45 -0800, SunWuKung wrote:
> > CREATE RULE new_entry AS ON INSERT to logview DO INSTEAD INSERT
> > (id,tm,info) VALUES (COALESCE(new.id,[default]),COALESCE(new.tm
> > ,[default]),COALESCE(new.info,[default]));
>
> what would [default] insert here?
> the default of the view or the default of the underlying table?
I admit, I haven't tested that myself yet.
But as I understand 'the theory', it would be the default of a VIEW.
This is because we don't have any means to access the [default] of an
underlaying table (nothing like 'foreign-default-reference' :), so we
define a [default] for a VIEW as an entirely separate entity.
This might be a problem for sequences, but in that case, we may put
*the*same* sequence for a VIEW [default] and for an underalaying TABLE
[default], so we don't actually need 'foreign-default-reference'
'construct' here. Other requirements for 'common [default]' might prove
more difficult to implement, but for me this is quite sufficient.
--
-R
From | Date | Subject | |
---|---|---|---|
Next Message | Marc Evans | 2006-12-14 12:58:45 | 8.2 server core dump |
Previous Message | Albe Laurenz | 2006-12-14 11:51:41 | Re: about the RULE system |