home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!sun-barr!cs.utexas.edu!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!ira.uka.de!math.fu-berlin.de!unidui!Germany.EU.net!mcsun!sunic!dkuug!diku!elgaard
- From: elgaard@diku.dk (Niels Elgaard Larsen)
- Newsgroups: comp.databases
- Subject: Re: Normalizing 2NF -> 3NF
- Message-ID: <1992Nov18.100829.3577@odin.diku.dk>
- Date: 18 Nov 92 10:08:29 GMT
- References: <1992Nov15.163413.17527@news.uiow> <49800010@ax.apc.org>
- Sender: elgaard@trud.diku.dk
- Organization: Department of Computer Science, U of Copenhagen
- Lines: 24
-
- villas@ax.apc.org writes:
-
-
- >Considering the normalization of (A*,B,C), where A -> B, B -> C and A* is
- >the PK, I would go with (A*,B) and (B*,C). My reasons are:
-
- > - If you go with (A*,B) and (A*,C), there no longer exists (or can be
- > implied) the B -> C dependencie, so the set of original dependencies
- > is being modified;
-
- Yes, you lose information (I didn't see that in my last posting).
-
- > - If you get to BCNF (Boyce-Codd Normal Form), you would also go with
- > (A*,B) and (B*,C), since every "determinant" must be a candidate key
- in the SAME relation
- > (in this case, they are the PKs in each relation), but that's not 3FN
- > (as a matter of fact, a step further 3FN).
-
- (A*,B) and (A*,C) are on BCNF
- --
- Niels Elgaard Larsen
- Institute of Computer Science,
- University of Copenhagen
- E-mail: elgaard@diku.dk
-