Re: 2nd Normal Form Question

From: Jan Hidders <hidders_at_gmail.com>
Date: Fri, 8 Feb 2008 10:30:13 -0800 (PST)
Message-ID: <1046f39d-c171-409f-9505-f539e7d8df08_at_m34g2000hsf.googlegroups.com>


On 8 feb, 17:55, gamehack <gameh..._at_gmail.com> wrote:
> On Feb 8, 4:39 pm, Jan Hidders <hidd..._at_gmail.com> wrote:
>
> > On 8 feb, 17:21, gamehack <gameh..._at_gmail.com> wrote:
>
> > [snip]
>
> > > But as we can see two non-prime
> > > attributes (party & home state) are functionally dependent only a
> > > subset of the primary key (the subset being the empty set). This must
> > > imply that it is not in 2NF.
>
> > They are not dependent on the empty set, but on the set {"Winner
> > Name"}, which is clearly not a proper subset of a candidate key. So no
> > 2NF violation there.
>
> But isn't it a requirement for 2NF that all non-prime attribs (party &
> home state) _have to be_ dependent on the whole of a candidate key (as
> we only have 1 candidate key here which is the primary key) which is
> not the case in this relation. Am I wrong?

Sorry, but yes, you are wrong. The attributes "Party" and "Home State" are both dependent on "Year". Note that "dependent on" means here that there is a functional dependency. It indeed holds that "Year"--
>"Party" and "Year"-->"Home State". This is not in contradiction with
the fact that it is also true that "Winner Name"-->"Party" and "Winner Name"-->"Home State". An attribute can be dependent on several different sets of attributes.

  • Jan Hidders
Received on Fri Feb 08 2008 - 19:30:13 CET

Original text of this message