Not sure how to map these particular types of relationships, and I can't find anything exactly like it after trawling the web - any help would be greatly appreciated. (Please note I've also gotten very little sleep as of late - so I might just be overlooking something)

They both revolve around audit tables where the primary key of each entity is a combination of the object's PKID + a "session" ID that created the records. Whenever the parent object is updated, all children are updated as well (regardless of whether they have changed or not), effectively giving them all the same "session" ID, and adding new records to the audit tables for both parent and child.

Parent
- id
- name
- age
- child_type_one_id
- session_id

ChildTypeOne
- child_type_one_id
- foo
- sessionId

ChildTypeTwo
- parent_id
- bar
- session_id

So when I load a particular version of a Parent from the audit table by (ID + SESSION_ID), how can I map the following two scenarios to have Hibernate load the correct ChildTypeOne and ChildTypeTwo records from those audit tables?

1. load the correct ChildTypeOne record, by using the Parent object's (CHILD_TYPE_ONE_ID + SESSION_ID) columns

2. load the set of ChildTypeTwo records, using the Parent object's (ID + SESSION_ID) columns

FYI - I'm using Hibernate 3.2

Thanks in advance.

1.



2.




Note childTypeOneId, sessionId and id in property-refs refer to properties in the parent class.


Farzad-

one another thing you could do for 1 is to use a many-to-one with two nested column tags. However, those two columns should match the primary key of childTypeOne which might not be want you want. Alternatively you might be able to use a join with a nested subselect to rectify this.


Farzad-