Hallo, i'm new to Hibernate.
For a new project i evalueted some O/R mapping tools and it seems that hibernate is one of the best. The object model exists and my part is to persist it.

Inheritance is used (1 base and 4 extended classes) and i need to map it as joined-subclasses with inverse="true"

Searching the forum i found this posting:
http://forum.hibernate.org/viewtopic.php?t=925839

I was able re-produce blip's problems - but where is the mistake?
When i change the inverse to "false" i got the correct entries in the database.

Can someone send me link to a working example or can help me getting the sample running?

Cheers,
Joey

I don't understand what blip is trying to suggest is wrong, so I'm amazed that you do. How about you tell us what your problem is, in terms that I will understand.

Here is my interpretation of the problem:

The 1:m ralation between the Person and the Vehicles is only stored correct in the database when the inverse setting is "false". When set to "true" the id_person is not updated in the vehicle table.

btw: I changed the "cars" one-to-many-class to test.Vehicle

These are my Testresults:

The first row is with the inverse="true"
The second row with inverse="false"

mysql> select * from vehicle;
+----------------------------------+----------------------------------+-----------+
| id_vehicle | id_person | sernum |
+----------------------------------+----------------------------------+-----------+
| 8a8a8bf8f91c5a7f00f91c5a822f0002 | NULL | 987654321 |
| 8a8a8bf8f91c5ad000f91c5ad2a20002 | 8a8a8bf8f91c5ad000f91c5ad28d0001 | 987654321 |
+----------------------------------+----------------------------------+-----------+

Could it be the bag parameter settings?

Show you whole mappings, please.

I hope this is what you want:

<?xml version="1.0"?>
<hibernate-mapping>
<class name="test.Person" table="person" dynamic-update="false" dynamic-insert="false">
<id name="personId" column="id_person" type="java.lang.String">
<generator class="uuid.hex"></generator>
</id>
<bag name="cars" lazy="false" inverse="true" cascade="all">
<key column="id_person"/>
<one-to-many class="test.Car"/>
</bag>
<property name="name" type="java.lang.String" update="true" insert="true" column="pname"/>
</class>
</hibernate-mapping>



<?xml version="1.0"?>
<hibernate-mapping>
<class name="test.Vehicle" table="vehicle" dynamic-update="false" dynamic-insert="false">
<id name="vehicleId" column="id_vehicle" type="java.lang.String">
<generator class="uuid.hex">
</generator>
</id>
<property name="serialNumber" type="java.lang.String" update="true" insert="true" column="sernum"/>
<many-to-one name="owner" class="test.Person" cascade="all" outer-join="false" update="true" insert="true" column="id_person"/>
<joined-subclass name="test.Car" table="cars" dynamic-update="false" dynamic-insert="false">
<key column="id_vehicle"/>
<property name="licensePlate" type="java.lang.String" update="true" insert="true" column="licplate"/>
</joined-subclass>
</class>
</hibernate-mapping>


P.S.: I don't want to prove the mapping is correct or not. I just want to see a running sample! Wouldn't it be good to put a sample into the FAQ?

And, when you add a Car to a Person, do you correctly set the backpointer by calling car.setOwner() ?

Well, no. I didn't know that i have to. I saw some one-to-many samples and it was never necessary. Is this a special "subclass" bahvior?

No, this is always required for bidirectional associations. Hibernate does not implement "managed associations" (for very good reasons).

Was my misunderstanding. I thought that the saveOrUpdate in combination with the cascade="all" and inverse="true" is the managed associations mechanism.

Now i am totally confused:

in Chapter 5.4 in the Hibernate documentation i found:



When i use blip's sample and set the id_person to NOT NULL and use inverse="true" the sample is only working when i use the "backpointer".

I don't understand why it is working with inverse="false". Am i stupid?
In this case the documentatin should give hints when backpointers are mandatory and when not.

backpointers are mandatory when you have a not-null constraint. (or any other time when inverse="true", I suppose).

Should i write a small FAQ for joined-subclasses?
It looks like that many other users have similar (understanding) problems.

This problem has nothing to do with <joined-subclass> mappings. Any other mappings behave exactly the same.