Re: History of 1NF
From: paul c <toledobysea_at_ac.ooyah>
Date: Thu, 13 Mar 2008 15:28:25 GMT
Message-ID: <tKbCj.84867$w94.1832@pd7urf2no>
>
> I don't think normal form depends on introducing more attributes, but on
> decomposing tables. The columns introduced, namely foriegn keys, refer to
> attributes that are already there.
Date: Thu, 13 Mar 2008 15:28:25 GMT
Message-ID: <tKbCj.84867$w94.1832@pd7urf2no>
David Cressey wrote:
> "paul c" <toledobysea_at_ac.ooyah> wrote in message
> news:1k%Bj.77466$w94.11329_at_pd7urf2no...
>> David Cressey wrote:
>
> I don't think normal form depends on introducing more attributes, but on
> decomposing tables. The columns introduced, namely foriegn keys, refer to
> attributes that are already there.
"Attribute introduction" seems part and parcel of the Information Principle and normalization a consequence of that. If you started from scratch wanting to know what is the most popular combinations of parts then without physical tricks, such as graphics, some data needs to be added to.
eg., customers A and B have both bought parts 1 and 2, customer C has bought part 1. It would be nice if a query answer didn't repeat customer numbers or pick one arbitrarily. Received on Thu Mar 13 2008 - 10:28:25 CDT