Re: WIP: Automatic view update rules - Mailing list pgsql-hackers

From Bernd Helmle
Subject Re: WIP: Automatic view update rules
Date
Msg-id E8182FB470BA1207E6A8838D@teje
Whole thread Raw
In response to Re: WIP: Automatic view update rules  ("Bernd Helmle" <bernd@oopsware.de>)
Responses Re: WIP: Automatic view update rules  (Bernd Helmle <mailings@oopsware.de>)
List pgsql-hackers
--On Sonntag, Dezember 28, 2008 15:29:58 +0100 Bernd Helmle 
<bernd@oopsware.de> wrote:

>> maybe the better solution is to not allow such a view to be updatable
>>
>
> Yes, it seems we have to check for target lists having negative attnums in
> checkTree(). Another solution would be to simply ignore those columns
> (extract them from the target list and include all updatable columns
> only).

While looking at this it turned out that the problem of updatability is 
more complex than
i originally thought: Consider a relation tree of the following 
views/relations:

View1 -> View2 -> View3 -> Table1

That means, View1 consists of View2 and so on. What happens now if someone 
is going to change View3, so that it's not updatable anymore? What the 
patch actually does is, scanning all relations/views involved in a current 
view (and cascading updates) und reject update rules as soon as it finds 
more than one relation within a view definition. Unfortunately this seems 
not to be enough, we had really check all involved views for updatability 
recursively. The infrastructure for this is already there, but i wonder if 
it could be made easier when we are going to maintain a separate 
is_updatable flag somewhere in the catalog, which would make checking the 
relation tree for updatability more easier.

Comments?

--  Thanks
                   Bernd


pgsql-hackers by date:

Previous
From: Heikki Linnakangas
Date:
Subject: Re: Visibility map and freezing
Next
From: Simon Riggs
Date:
Subject: Re: Hot standby, slot ids and stuff