I probably don't understand the complexities with this, but wouldn't the ID values in AA have to be completely independent of the values in A?
> On 28 févr. 2012, at 17:59, Steve Ebersole wrote: > >> Another discrepancy between hbm.xml and annotations currently is the >> definition of natural ids. hbm.xml only allows natural ids to be >> defined on the root entity, annotations allow it to be defined on any >> level of the hierarchy, even across classes. >> >> We need to decide what we want to support. 2 things to consider here are: >> 1) whether we limit @NaturalId to only root entity meta (and maybe >> @MappedSuperclass for root entities) >> 2) if we do not limit, what @NaturalId spread across a hierarchy "means" >> >> In the second point, consider: >> class A >> @NaturalId >> key1 >> >> class AA extends A >> @NaturalId >> key2 >> >> Personally I vote for limiting @NaturalId to the root entity. This >> allows caching to work just like it does for identifiers. JPAV _______________________________________________ hibernate-dev mailing list hibernate-dev@lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev