Re: Constraints and Functional Dependencies

Walt wrote:
> "Marshall" <marshall.spight_at_gmail.com> wrote in message
> news:1172382132.346588.167980_at_s48g2000cws.googlegroups.com...
...

>>>In practice, it turns out to make data management easier if,  wherever

>>>are tables with multiple candidate keys, one of them is chosen as the so
>>>called "primary key",  and all foreign key references to that table are

>>>by way of that key.
>>
>>Can you list a few ways primary keys make things easier?

As John Lennon said, "You know my name, look up my number". Just to drop a couple more names, I once heard Codd speak for an hour about nothing except Christie Brinkley's phone number. I'm quite sure what he called primary keys and other people call candidate keys were part of his theory, just as they are part of my understanding of that theory.

They are the most fundamental way to make inferences. Fundamental because if the computer is ignored one can infer from them more or less without any relational operators. In fact, that was true before Codd.

I'll grant that a system could get the same answers without keys, but it would take a lot more effort and a lot more plumbing. If we didn't have them, I'd like to know how we would ensure that a person doesn't have more than one name. The same ways that people did that with IMS or IDMS? Uggh!

p Received on Mon Feb 26 2007 - 23:10:53 CET