Alfred Perlstein wrote:
> Thank you, it's not a big problem that this doesn't happen, but it'd
> be nice to see it as an option when creating a table via inheritance.
>
> What about RULEs? I wouldn't really have a use for that but others
> might.
Actually it's a reasonably big deal. Apart from the obvious performance
penalty for a deep inheritance hierarchy it affects the implementation
of unique keys, referential integrity for inheritance and the
transparancy of extending an inheritance hierarchy.