Functional Dependancies and Normalization

From: Mr B <craig_at_mrb.u-net.com>
Date: Sun, 27 Aug 2000 19:54:09 +0100
Message-ID: <bdoiqso36504l5silt77743ng0fpkgn3ou_at_4ax.com>

Can someone give me a few pointers here...

I am studying databases and am a bit confused about reverse functional dependencies when used in the mormalisation process. I understand that there existence depicts a 1-1 realtionship and that the normalisation process cannot produce 1-1 realtionships.

My question is...what do you do wqith a reverse FD before normalising as its existence can be used to produce derived FD's. Do these derived FD's still hold ? Do you exclude the FD at the start ?? If so which one do you exclude (obviously there are two i.e. A->B and B->A).

Some guidance on this or references to text about this would be greatly appreciated.

Thanks,

Craig.

---

Craig Blaikie
Scotland.
e-mail me at:- craig.blaikie_at_ntlworld.com Received on Sun Aug 27 2000 - 20:54:09 CEST