>
> Actually, the information_schema supposes that constraint names are
> unique within a *schema*, not within a *catalog* (a/k/a database).
> Don't know if that distinction can help you or not. You are correct
> that Postgres is less rigid. We do not consider that to be a deficiency
> on the Postgres side ;-)
Yes, my bad. Anyway, it doesn't help, I have both tables on the same schema.
>
> If you want to use the information_schema to deal with this stuff, the
> answer is to make sure that your application follows the SQL-standard
> rule of not duplicating constraint names within a schema.
>
> regards, tom lane
>
Yes, I know that following the SQL standards is the way to go, but
sometimes this has to be done in databases I don't design, so I have
to be prepared for every case. I think I'll use the pg_catalog for
this case.
Thank you for you answer,
Jonathan
