Thanks for the feedback.

Actually we use a native query through JPA/hibernate.

Regarding the Dozer way I'm unsure how "generic" it would be: we're doing this for an existing project and we wanted to avoid doing per entity tuning.

Anyway, the main point is that it feels like a pretty common use case and doing it through native queries feels awkward.

We had to fetch each table definition to know how to display it's content. From envers itself this could be obtained through the entities definition directly, thus saving this database roundtrip.

If we were to move to partitioned tables (we should!), then we'll have to tweak our SQL again.