--Jan:
--don't worry, i wasn't offended; i'm sorry
i didn't wait around for someone to point
me in the right direction.
--and messing around in a system table
that i have very little knowledge of scared
me, too ... i *wanted* to be discouraged.
--what motivated me was something i read
in the groups.google.com comp.databases.postgresql.admin
news group. someone had a similar problem and it was
pointed out (me thinks by tom lane?) that
there could be bogus value in pg_class,
pg_function, etc ...
--i deleted the function and then the rows
where the typowner was '0' and the names typname
was '<whatever>'.
--from now on i will wait until i get a better
handle on how i butchered the db before i go
guns-a-blazing.
--thanks for the follow up!!
-X
-----Original Message-----
From: Jan Wieck [mailto:janwieck@yahoo.com]
Shaunn,
that reply below wasn't meant personal or in any way
offending. I more wanted to use it as a chance to discourage
people from manipulating the system catalog in general. This
is usually a bad idea and since system catalog structures can
change from version to version, what worked perfectly
yesterday could end in a catastrophe tomorrow.
Jan
Jan Wieck wrote:
> Johnson, Shaunn wrote:
> > --Never mind; I solved it.
> >
> > --It had to do with something in pg_type.
>
> Just deleted the rows from pg_type?
>
> If so, congrats, in this case it is one of the possible
> solutions. In some other cases you might well mess up your
> entire DB's system catalog by shooting around in the pg_*
> tables like that without the advice of someone who really
> knows what she's doing there.
>
>
> Jan
>
