Code generation for associations

This seems to be a bug in SDE, since VP generates arrays for such associations. It’s better but I’d like to be able to choose between arrays or collections.

Hi Jean-Marie,

Thanks for your suggestion. The next release shall be able to generate an array. The reason why it will not generate a collection (such as a Vector) is that the type is unknown.

We suppose that, a collection shall be supported when JDK 1.5 is released as it supports generics.

Hope this helps.


Hi Martyn,

Thanks for your fast reply, I apologize for the many posts concerning this question.

You’re right: I wrote “collection” but thought about generics :wink:

Nevertheless, I’d like to know if there is a bug in SDE, which doesn’t generate an array, whereas VP does. For the time being, I’l simply disable the code generation for associations, this is not a really big problem.

Again, I like SDE and will probably choose it for my team.


Hi Jean-Marie,

It’s not really a bug as such, we just did not implement it for the SDE’s. It will be implemented in the next release which should be quite soon.

Sorry for the inconvenience.



I have one more small problem with SDE (which I still like very much): I can’t see the Code Generation Options dialog box, which is present in Visual Paradigm and mentionned in the Help (which is simply the VP documentation).

It’s important to be able to keep SDE from generating code for n-ary associations, since it generates simple member variables instead of collections or generics (see previous messages in this thread).

Every time I manually change for example :
private Foo <---- Code generated for the n-ary association
to :
private Foo[]

in the code,

  1. Reverse engineering breaks the relationship (she looses its name)
  2. Forward engineering still generates ann unuseful member variable.

Thanks in advance for your replies, I’m apologize if this is a FAQ.


I’m currently evaluating SDE for Intellij IDEA, build 20040414a. I’m very impressed by this tool, of course.

But I’m having trouble with code generation for associations: a 1-to-* association is translated as a 1-to-1 association, which is clearly not what we expect. I’d prefer getting a Collection or an array, or be able to choose, of course. In reverse engineering, I get the same problem: this kind of associations are not recognized.

This is a silly question for sure, but thanks for answering it :wink: I “fought” with this problem for a while but found no solution…