I poked a bit further and looked broken dependancies for views, PL/pgSQL
functions and triggers. It looks like PostgreSQL already handles the
trigger dependancy correctly. The function relation/OID resolution is
cached for the lifetime of the backend process. Once I quit and
reconnected it reresolved the table name again. The issue remained for the
view object.
So... A function could be created to drop the view and recreate it from
the saved source function.
Joshua b. Jore ; http://www.greentechnologist.org
On Sun, 7 Jul 2002, Terry Yapt wrote:
> Only a note:
> ===========
>
> Oracle do it without any problem... :-) So because of that It is something normal for me...
> In oracle you can looking for INVALID objects... In our example, function 'x' will be an invalid object..
>
> Perhaps it is the solution for postgreSQL..
>
> Thanks Josh and best regards..
>
>
> Josh Jore wrote:
> >
> > You examine the oid attribute in the pg_class system table (from
> > [hackers], this might not be valid advice past 7.3). This assumes you
> > looked at the oid *before* deleting the table. This behaviour doesn't mean
> > you have to track which things have which oid - what you suggested. It
> > does mean you have to know that the function "x"(..,..) depends on the
> > table "pepe". You should already know your application's dependancies
> > to begin with. This isn't rocket science - either your developer documents
> > the design or you inspect the source.
> >
> > Now... it would be *nice* if the symbols could be re-resolved as needed
> > either as a manual or automatic process. I can dream can't I?
> >
> > Joshua b. Jore ; http://www.greentechnologist.org
> >
> > On Sat, 6 Jul 2002, Terry Yapt wrote:
> >
> > > Ok, I understand that (I imagined this behaviour also) but how can
> > > locate what object oid is '16561' to re-create it ? Are there any
> > > system table to do it ?
> > >
> > > And why 'Select x(..,..) runs in pgadmin and not in psql ???????
> > >
> > > I think if I must to maintain an, external to pgsql, document with all
> > > my objects references noted then I think pgsql development team have a
> > > lot of work to do yet. :-(
> > >
> > > Thanks Josh, best regards..
> > >
> > > PS: amusing self_confidence.jpg <g>.
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org
>
>
>
