Re: By The Dawn's Normal Light
From: Paul <paul_at_test.com>
Date: Fri, 22 Oct 2004 23:59:54 +0100
Message-ID: <417990eb$0$47999$ed2e19e4_at_ptn-nntp-reader04.plus.net>
>
> I found a paper that I bought from Date last year specfically on 1st normal
> form. He starts with a quote from Codd
>
> "A relation is in first normal form if ...none of its domains has elements
> that are themselves sets. An unnormalized relation is one that is not in
> first normal form."
Date: Fri, 22 Oct 2004 23:59:54 +0100
Message-ID: <417990eb$0$47999$ed2e19e4_at_ptn-nntp-reader04.plus.net>
Dawn M. Wolthuis wrote:
>>1NF is superfluous.
>
> I found a paper that I bought from Date last year specfically on 1st normal
> form. He starts with a quote from Codd
>
> "A relation is in first normal form if ...none of its domains has elements
> that are themselves sets. An unnormalized relation is one that is not in
> first normal form."
I'd say that 1NF is different from all the other NFs because it is an attribute of the DBMS rather than an attribute of the table.
The requirement that none of the domains have elements that are sets needs clarifying I think by saying that this is from the perspective of the relational engine. So even if the elements are sets, this is hidden, except to the type engine.
You could define the integers as sets, for example, as a set foundation of mathematics would do, starting with the empty set. But this doesn't mean that a relation can't have integer domains.
Paul. Received on Sat Oct 23 2004 - 00:59:54 CEST