Re: BCNF: superkey or candidate key ?
Date: Mon, 25 Sep 2006 19:36:14 GMT
> Yes it would, but the determinant in B -> X is not a candidate key, it can
> be reduced, it's a superkey.
> Maybe the answer is that considering the closure of FDs, we find that
> candidate key is too strong a term and that superkey is now accurate.
If one added the additional constraint that B -> X is left-irreducible, would any proper superkeys remain to consider? Received on Mon Sep 25 2006 - 21:36:14 CEST