Re: Stored fields ordered left to right
Date: Mon, 12 Jan 2004 04:28:56 -0600
Message-ID: <bttstc$mda$1_at_news.netins.net>
"Adrian Kubala" <adrian_at_sixfingeredman.net> wrote in message
news:slrnc04g10.qpa.adrian_at_sixfingeredman.net...
> Dawn M. Wolthuis <dwolt_at_tincat-group.com> schrieb:
> > Every relation where each tuple has a unique identifier or candidate key
can
> > also be represented as a function. Can you give an example of a set of
> > propositions that can be modeled as relations but not as functions?
>
> {<1, 1> <1, 2> <2, 1>}
Yeah, thanks Adrian, but I've got the mathematics part down on the difference between a function and a relation. I can graph such. But when it comes to what we once called data processing using a computer, is it feasible to have a pure relational implementation (no, I'm not confusing implementations with the model -- stick with me) using a database without actually ending up with a function? In the relational model, how do you implement this without either a) an identifier tagged on by the database which could be kept out of the model or b) a candidate key?
The function this would correspond to in the model I work with is one where either the designer would implement it with a random or sequential key, such as:
ID: 1
INFO: <1 ,1>
ID: 2
INFO: <1, 2>
ID: 3
INFO: <2, 1>
Or perhaps:
ID: 1
ID:<1,1>
INFO-A: 1
INFO-B: 1
ID: 2
INFO-A: 1
INFO-B: 2
ID: 3
INFO-A: 2
INFO-B:1
You can see the function outright in the above design.
Or the designer could add it without such a key, by making the entire tuple
a "candidate key" by implementing it as
ID:<1,2>
ID<2,1>
In this design, the function maps each of these IDs to the null set.
I might have said that poorly -- did that make sense? Thanks a bunch -- I truly do want to understand and be corrected if I have this wrong. --dawn In Received on Mon Jan 12 2004 - 11:28:56 CET