cayenne 3.0.1 reverse engineering on IBM iseries db2 tables, not identifying primary keys

["Joseph P. Berglund" <jo...@yahoo.com>]

Cayenne Users,

I am evaluating whether I can use Cayenne 3.0.1 for my company, we use an IBM iSeries DB2 (as opposed to DB2 universal database). The cayenne modeller reverse engineering tool correctly identifies the schemas, attributes, and their types but fails to identify the primary keys and indexes.

The primary keys are simply not marked as being primary keys. This would normally be no problem, but I am dealing with hundreds of tables and it would be a pain to have to look up the PKs and mark them all.

Furthermore on a minor note, the indexes are being added as tables (also without PKs). I-Series DB2 does intentionally expose indexes as tables to JDBC, but I was hoping the adapter would know the difference (they really are unnecessary as the database should decide what index to use based on the query..)

If you can tell me if there is something I am missing to be able to have the reverse engineering tool identify the primary keys on db2, it would help me out a ton. I have used Cayenne before with MySQL 5.1 and have had no problems with reverse engineering. 

My environment:
I am using the JTOpen db2 driver (which is recommended by IBM) and the Cayenne DB2 adapter. The tables are hosted on an IBM v6r1 installation with DB2 (DB2 is paired with the OS). My client is a windows xp sp3 machine with an IBM Websphere 6.1 servlet container, production is a WebSphere hosted on the same iSeries as the tables.

Thanks,
joey b.