Hi,

i m not sure if i m right in this group, but i think relational
database modelling is a theoretical issue:

I have to following (simplified) situation. I have a company which
sells

- simply one article
- a basket with an amount of the same article
- a customer special bill of material (n articles with amount)

Now the question is: How should i correctly model that with sql on a
relational database?

3 different tables for my products? But than: What should i reference
in the sales document (also a table)?

Constraints:
- It s not allowed to use positions for the basket
- It s not allowed to add a foreign key (to the sales document) to the
article table

In my opinion, this problem is prety much a object orientated
inheritance problem ...


Any help appreciated,
Dan

Daniel Austria wrote:

Sorry, Dan, this is not comp.databases.do.my.homework.for.me

hi thanks anyway, i found a cool link:
http://jelle.druyts.net/2004/04/14/I...lDatabase.aspx

Daniel Austria wrote:

I didn't find anything cool about it. I found it very poor in fact.

Daniel Austria wrote:

Not really. You're not here to discuss theory,
but to apply it.

One table for the articles.
No table for baskets, as far as I can see.
One table for customers.
One table for orders, with columns referring to customers and articles,
and another column for the amount.
You probably want multiple rows in this table pertaining to the same order.
Not sure whether you want some kind of order number to identify the order,
you may not need that.

I think it's a selection from a table: the rows in the orders table that
satisfy certain criteria (having the same order number, or the same
customer and order timestamp, or somethign else, depending on how you
identify them).

Why not?

I don't see how.

--
Reinier

Hi Reinier,

thanks for your answer. In the meantime i found a complete description
and solution of my problem in

Martin Fowler, Patterns of Enterprise Application Architecture

There is a a chapter Structural Mapping Patterns (41 ff.), which
discusses mapping of compositional hierarchies to relational
databases.

Best,
Dan

Daniel Austria wrote:

Fowler is certainly better than the 1st reference you cited, but I think
you can do a lot better even still.