3NF but not BCNF

From: Phil Cook <pacookSPAM_IRIS_at_blueyonder.co.uk>
Date: Tue, 15 May 2001 10:09:41 GMT
Message-ID: <Ff7M6.8074$zq2.600444_at_news1.cableinet.net>


I am presented with the following question:

    Argue that if a relation schema R is in Third Normal Form but not in Boyce-Codd

    Normal Form with respect to a set of functional dependencies F, then it must have

    at least two distinct keys for R with respect to F which overlap, i.e. such that their

    intersection is nonempty.

Unfortunately, my textbook only mentions this point in passing, referring to some paper by Vincent and Srinivasan. This paper does not appear to be available online, however.

Any explanation would be greatly appreciated. Received on Tue May 15 2001 - 12:09:41 CEST

Original text of this message