Thank you, that may be obvious but again I'm a newbie! If there's an obvious resource I should be accessing please just point me in th right direction.
In fact I don't even know how to get a hold of EA support. If that requires licensing, perhaps I can expedite processing through Purchasing.
Anyhow, what I'm finding is that once I use the <<table>> stereotype I can generate DDL but no longer Java. This is true even if I turn the stereotype off, there must be some marker remaining in the metadata.
This is really a problem since I need to manage entity bean, database and XML-Schema off a common model. I need to do some extensive refactoring of the current database design, splitting and merging many tables.
Not only can't I configure the model to generate both DDL and JavaBeans, but I don't even see a straightforward way to maintain two versions of the model. For example Rational Rose has automated facility to create a deployment model and traces from an analysis model, but I don't see how to do that from EA. I don't even see a way to copy a collection of objects from one package to another, but again I'm a newbie.
Can anyone tell me how I can unmark my objects so I can generate Java again? Perhaps I can create my database, reverse engineer the tables, then change the tables to plain old Java objects.
While we're at it, the package statements in the generated Java reflect the entire package stucture e.g., "Data Model.com.psi.e_and_e.domain" even when I select com as the root directory. Is there any way to generate Java properly from a subfolder?
Returning to the common model issue:
Perhaps I should generate my own schema etc off the back end by postprocessing the XMI. Are there an API and examples for programming against the metadata?
Or perhaps I should insert XDOCLET metadata in the comments and use that to generate beans. Has anyone tried this?
Surely we're not the first ones to try modelling entity EJB with EA!
