Re: NULLs

On 5 jan, 16:29, "David Cressey" <cresse..._at_verizon.net> wrote:
> "Roy Hann" <specia..._at_processed.almost.meat> wrote in message
>
> news:h6adnSi6_pWI5eLanZ2dnUVZ8sijnZ2d_at_pipex.net...
>
> > "David Cressey" <cresse..._at_verizon.net> wrote in message
> >news:VwKfj.321$vX6.267_at_trndny05...
>
> > > I disagree. A null doesn't indicate that information exists.  There is
> > > always the "not applicable" case
>
> > Which, I believe, is almost invariably the reason that nullable attributes
> > get introduced in real databases.  I continue to be baffled at this
> > preoccupation with flagging information that must exist but isn't known.
> > Discuss it if you must (I absolutely don't care about it), but don't
> ignore
> > the more common problem.
>
> The "not applicable" case is the one that can be obviated by normalization.

That is indeed the obvious case.

> Even with full
> normalization there can be facts that the database doesn't know, and knows
> that it doesn't know.  There can also be facts that the database doesn't
> know, and doesn't know that it doesn't know.

Here is a better explanation:
http://www.youtube.com/watch?v=jtkUO8NpI84&feature=related

:-)
Sorry, could not resist.

Regards,
Carlos Received on Sun Jan 06 2008 - 02:14:37 CET