Re: two nasty schemata, union types and surrogate keys
Date: Sun, 27 Sep 2009 14:01:14 GMT
> The complement of a union is exactly those tuples that can
> theoretically appear in either operand but don't; the complement of a
> join is exactly those tuples that can theoretically appear in the
> result but don't.
Why put it differently for union as compared to join? Surely the complement of a union is exactly those tuples that can satisfy the result but don't. I would say that way of putting it makes it more obvious which predicates can be satisfied by various tuples and which predicates can't, likewise which propositions are possible and not possible. Received on Sun Sep 27 2009 - 09:01:14 CDT