Hi,

Some simple questions about how does hibernate (3.2.6/oracle) handling optimistic version control with one-to-many model constructions.

Consider following:

I have simple one-to-many relation (without any cascades defined) where both parent and related MOs have optimistic version control defined on (with "version" annotation). When I initialize both parent and collection of it's related objects in session (smth. like that: find("FROM Parent p JOIN FETCH p.childs")) and update only and only some child entity from related childs collection (some object from the many-side collection/childs) then hibernate automatically performs SQL update both for updated entity from parent's child collection (that was actually updated) as well for parent (to increase it's sys version field by 1).

Why does it update parent when it was unchanged and there is not any cascade defined as well? How can I prevent this or to get rid of inefficient implementation like this?

I found link to issue and as far as I concern this is defined as "normal" strategy in hibernate?
http://opensource.atlassian.com/project ... se/HB-1514

Thnks,
Sincerely
Ilhan

Hi,

Answer is that this default feature is connected with @OptimisticLock(excluded=true) configuration option.

Thkns.
I.