Tom Lane escribió:
> Oleg Serov <serovov@gmail.com> writes:
> > So there are no simple way to do it right, and it will be not fixed? Will
> > this bug appear in todo list?
>
> It's not a bug, it's just what happens when you make the parent and
> child column orders inconsistent. Would you prefer that we restricted
> ALTER TABLE to refuse to perform the alteration in the first place?
Possibly the right fix is to implement the oft-discussed idea about
columns able to be moved around in table definitions. That would let
ALTER TABLE ADD COLUMN to put the column in the right positions in
children.
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support