Re: Newbie question about db normalization theory: redundant keys OK?

From: David Portas <REMOVE_BEFORE_REPLYING_dportas_at_acm.org>
Date: Fri, 14 Dec 2007 07:10:07 -0800 (PST)
Message-ID: <896c850b-225c-4c1c-bede-e89ef5b63b25_at_s8g2000prg.googlegroups.com>


On 14 Dec, 13:07, "David Cressey" <cresse..._at_verizon.net> wrote:
>
> A few comments here:
>
> 2NF applies only to compound keys. A table that is in 1NF and has only a
> simple key is in 2NF necessarily.

That's incorrect, as discussed previously: http://groups.google.com/group/comp.databases.theory/browse_frm/thread/86b907799dfb71f6/0cf0bdb0e48770b0

> 1NF, 2NF, and 3NF are not the end of the story. In addition, there are
> BCNF, 4NF, 5NF, and a final normal form called domain-key normal form.

In what way is DKNF "final"? Not in any sense that I can see. A DKNF schema can still be non-loss decomposed. In fact DKNF is not like other NFs at all since it prohibits you from expressing schemas with certain constraints and so presumably can be applied in special cases only.

> I've een references to 6NF in this newsgroup. As far as a can tell, 6NF is
> the same thing as domain-key normal form.
>

It is not.

--
David Portas
Received on Fri Dec 14 2007 - 16:10:07 CET

Original text of this message