Re: foundations of relational theory? - some references for the truly starving
Date: Sat, 25 Oct 2003 19:15:21 -0400
Message-ID: <GumdnaAjFvrJmQaiU-KYvg_at_golden.net>
"Marshall Spight" <mspight_at_dnai.com> wrote in message news:8vxmb.25533$e01.51735_at_attbi_s02...
> "Bob Badour" <bbadour_at_golden.net> wrote in message
news:0c6cnW994f01HgeiXTWJkQ_at_golden.net...
> > >
> > > BTW, normalization is the process of removing redundancy
> > > from a schema. A fully-normalized schema is said to have
> > > no redundancy, not be fully two dimensional.
> >
> > A fully-normalized schema can have considerable redundancy. It just has
no
> > redundancy that one can remove with lossless decomposition using
project.
>
> Hmmm. The fact that I don't completely understand that
> leads me pretty strongly to believe that I don't understand
> normalization well enough. Does this imply that there
> is redundancy that can be losslessly removed using
> operations besides project?
It implies there are other operations by which one can losslessly decompose a relation. Whether those decompositions will reduce redundancy is another matter.
The values in a foreign key reference are redundant because they appear in multiple relations. In this case, the redundancy is appropriate and necessary to represent the data. Derivable attributes are also redundant. Thus, in a database of drivers in the province of Ontario, one might use the Ontario Driver's License Number as the primary key of some relation that contains the driver's surname and the driver's sex. However, one can derive the driver's surname and the driver's sex from the license number.
In this sense, every view and every snapshot is redundant.
> (Otherwise, I don't see how
> there can be a claim of redundancy.) What might
> those operations be?
There might be no operations by which one can remove the redundancy through any lossless decomposition.
> If you have further suggested reading, I would appreciate it.
> I'll also be on the lookout myself.
Date and Pascal have mentioned the issue a number of times such as in Date's reply in http://www.dbdebunk.com/page/page/622144.htm Fabian mentions a chapter in his latest book at the end of the article as another reference. Received on Sun Oct 26 2003 - 01:15:21 CEST