Re: Issue with functions in Rule

Hello Richard,

thanks for your response.
Your testcase does indeed make no problems.
I stripped down my case to be as simple as yours
and it worked too.

Now i try to put the things in it, that are usefull
for me (Validation in the update-function and Conversion
in the view).

I hope (or better not :~))i can then put a testcase together,
that fails on a specific point.

I'll be back with more information.

Best regards

Hakan Kocaman

Software-Developer
digame.de GmbH
Richard-Byrd-Str. 4-8
50829 Köln

Tel.: +49 (0) 221 59 68 88 31
Fax: +49 (0) 221 59 68 88 98

Email: hakan(dot)kocaman(at)digame(dot)de

> -----Original Message-----
> From: Richard Huxton [mailto:dev(at)archonet(dot)com]
> Sent: Wednesday, October 19, 2005 9:59 AM
> To: Hakan Kocaman
> Cc: pgsql-general(at)postgresql(dot)org
> Subject: Re: [GENERAL] Issue with functions in Rule
>
>
> Hakan Kocaman wrote:
> > Hello Folks,
> >
> > first i want to apologize for my bad english :~)
> > we got here apparntly 2 issues with the above mentioned
> functionality:
> >
> > 1. I want to use functions with composite types as parameters,
> > which is no deal on our current production
> server(postgres 7.4.3)
> > but don't work on our upcoming production server(8.0.3).
> > A sample function is attached.
> > The function uses a composite type from a view, which
> declaration
> > is alos attached.
> > In this thread it is mentioned as a bug:
> > http://archives.postgresql.org/pgsql-admin/2005-08/msg00153.php
> > The question is if it's fixed in 8.0.4 or 8.1.
> >
> > 2. Using the mentioned function in a rule on the mentioned view
> > clashs on the fact, that we don't find a appropriate
> way for the syntax.
> > The rule is also attached.
>
> CREATE OR REPLACE RULE update_produkte AS
> ON UPDATE TO viewprodukte DO INSTEAD
> SELECT fu_upd_viewprodukte(vp1.*, vp2.*) AS fu_upd_viewprodukte
> FROM viewprodukte vp1, viewprodukte vp2
> WHERE vp1.id = new.id AND vp2.id = old.id;
>
> Well, as a short-term workaround, surely you could rewrite
> this to take
> (new.id,old.id) as parameters instead?
>
> I also can't reproduce your problem in 8.0.3 on my
> Debian-based system
> here. I've attached my test-script - am I missing something?
>
> --
> Richard Huxton
> Archonet Ltd
>