prosper2000 wrote:
The table structure is also as above, with a parentAId and a childAId. The B tables share the same key (parentAId or childAId respectively), as current hibernate requires.
Same value is required, not same name.
prosper2000 wrote:
The code does the following - with each access to hibernate in a separate session:
?? what do you mean by each access in a separate session. Show the exact code, it's easier to understand than you five points.
prosper2000 wrote:
second, why is the save of the ChildBs, when Parent B cascades, not also causing duplicate ChildAs?
Probably related to save != saveOrUpdate because of unsaved-value