Data Modelling Strategies

[brucef]

I have just discovered the data modelling features (<<table>> stereotype etc). I am trying to design a process which has a fairly close meshing with a relational database and this is excellent.

What I am trying to work out is how my <<table>> classes relate to the rest of my design. I had planned a class to correspond to each table, with rows mapping to object instances. I can generate the SQL to create the tables from the <<table>> classes, but I can also generate C++ classes from the tables. Should I use these generated classes? Should I add the extra functionality to these classes, or derive subclasses, or perhaps the class should have some sort of relation with the table class.

I am sure this is a common sort of problem, perhaps people could describe how they go about modelling their data-oriented designs?

[CJ]

G'day,

Check out this article:

http://methodsandtools.com/archive/archive.php?id=9

Sorry, I don't have any better answer than that ... still trying to figure things out myself.

[brucef]

This is exactly what I was looking for, thankyou. I wonder why there isn't a copy of the document, or a pointer to it, somewhere on the sparx site? Perhaps there is and I missed it.

[Eve]

I can't find a link to it anywhere on the Sparx Systems website, but I could find a copy of it in google.

http://www.sparxsystems.com.au/uml_topics/uml_datamodel/uml_datamodel.htm

It seems to have diagrams that the other one doesn't.

[CJ]

G'day g'day,

Thanks Simon.  

Is that link on your web site?  If not, your UML Tutorials page would be a great spot.