Ferdinand Gassauer wrote:
> it would be great to have a coalesce2 function which treats empty strings as
> null values.
I disagree, it would be the opposite of "great". "" is a /known/ value and not
in the same semantic space as NULL at all.
If you really feel you need such a function, though, why don't you write one?
I recommend sticking with the SQL semantics instead of trying to change them.
-- Lew