Re: Modeling Address using Relational Theory

Mike wrote:
> On 26 Aug 2005 20:58:04 -0700, "dawn" <dawnwolthuis_at_gmail.com> wrote:
>
> [snip]
> >What I keep seeing in self-described relational models is Line1, Line2,
> >Line3. That just seems wrong. What would be the purest relational
> >approach to this?
>
> To express our understanding of addresses in the real world in terms of
> a logical data model. In other words, I'd do a conceptual model first,

do you have any suggestions for what entities, attributes, and relationships you would put in a conceptual model for an address?

Perhaps I can get that started -- an address is an identifier for a location (removing the name of a party that would show up in a mailing address from the address to be modeled). The attributes of a location are that it is in a political-geographic region, such as city, country, state or province. Additionally, the country has assigned a postcode for this region. In addition, ... (that's the part I need you to fill in) Thanks in advance if you do so!

> then translate it into a relational model. Gaining the understanding
> you need about how addresses work in the real world is beyond the scope
> of the relational model.

yup

> I get the feeling that a lot of people in this thread think an address
> and a mailing label are the same thing. They aren't.

agreed. Thanks. --dawn

>
> --
> Mike Sherrill
Received on Tue Sep 06 2005 - 20:09:20 CEST