Re: Pizza Example

From: Dawn M. Wolthuis <dwolt_at_tincat-group.com>
Date: Thu, 8 Apr 2004 16:16:09 -0500
Message-ID: <c54ff7$d46$1_at_news.netins.net>


"Jan Hidders" <jan.hidders_at_REMOVETHIS.pandora.be> wrote in message news:h2jdc.64877$%m6.4324481_at_phobos.telenet-ops.be...
> Dawn M. Wolthuis wrote:
> >
> > So, if "the Pizza has Mozzarella and Parmesan cheese" is stored as:
> >
> > Pizza Mozzarella
> > Parmesan
> >
> > in a single, uh, record, then if the ordering is unwittingly useful, we
> > didn't harm that ordering.
>
> Indeed, but you did make certain query-optimizations impossible.

If there were emperical data suggesting that queries on this model in PICK were faster than those on a 1NF model in any DBMS purporting to be relational, then I would not be concerned. For the purposes of this discussion I'll assume there is sufficient evidence to suggest that we are not losing anything in performance. However, I can imagine (but don't know) that parallel processing of queries could have a chance at speeding up relational implementations and leaving this one behind.

> Never
> mind that what you are advocating is basically sloppy data modelling.

Oh, please explain. I would suggest that if the user suggested from the start that they were only interested in at most one Nick Name, then the Nick Name would have ended up in the same relation as the First & Last names. Then when a different user years later says they want multiple Nick Names, we just allow for a list, rather than a single value, as the value of "Nick Name" and we are golden. So, if the requirements stated that we needed at most 1 Nick Name per person, then would it be sloppy to model it this ay? --dawn

> -- Jan Hidders
Received on Thu Apr 08 2004 - 23:16:09 CEST

Original text of this message