if I use primary keys to access Entity Beans, I can avoid all database accesses (using commit option A) since the cache is addressed by the primary key. But how could I tell JBoss to handle secondary keys (unique keys which could be also used as primary keys) in the same way? If we could avoid database access here this would be a great performance boost for us.
Currently I have two ideas:
1. Define a second Entity Bean with the secondary key as PK and which accesses the same table. Does this help to avoid database accesses? Or do I need to change the commit option?
2. Use CMR: Is it possible to define a relationship of a bean to itself? How could I do that? And what role plays the foreign-key attribute of CMR here?