Re: 2NF and 3NF obsolete ???
Date: Fri, 30 Nov 2001 22:09:01 -0500
Message-ID: <3C0849CD.CB51DAE1_at_ift.ulaval.ca>
Hi Frederic,
The problem, to pursue your logic will be to prove that no others FD but
those with a key on left side exists when the key is a composite one. If
all of your keys are made up of one attribute, then there is no problem
iff you can prove that no others complex keys does exists. To do so, you
have to rely on covert of FD.
AG
Frederic wrote:
> I'm preparing a course on normalisation and normal forms.
> I am actually consulting a book from Thomas Connolly et al.
> where it is stated that a table is in BCNF iff:
>
> Every determinant of the table is a candidate key.
>
> Well, I guess the table may also need to be in 1NF
> even if the authors do not mention it. Anyway this is
> not the point of my message.
>
> What strikes me is that it is very straitforward to
> verify this condition and also very easy to normalize
> afterward.
>
> Why bother anymore with 2NF and 3NF ?
>
> Frederic
Received on Sat Dec 01 2001 - 04:09:01 CET