Hello,

When class contains a one-to-many relation with a cascade attribute, objects are inserted in three stages. First insert is performed for the parent object, after that all child objects and at the end all child objects updated with their parent id.

For example (see mapping below) when storing A that has a reference to B instances. First A is inserted, then all Bs are inserted and then all Bs are updated with A's id.

The question is how (if possible at all) to exclude updates executed on Bs?

Thanks,

Alex

--- Info:

Hibernate version: 3.0.5

Mapping documents:

<hibernate-mapping>

<!-- class A -->

<class table="A_TAB" name="myPackage.A">
<id type="java.lang.String" column="ID" length="32" unsaved-value="null" name="id">
<generator class="uuid.hex"/>
</id>
<set fetch="join" lazy="false" cascade="all-delete-orphan" name="attributes">
<key column="B_ID"/>
<one-to-many class="myPackage.B"/>
</set>
</class>

<!-- class B-->

<class table="B_TAB" name="myPackage.B">
<id type="java.lang.String" column="ID" length="32" unsaved-value="null" name="id">
<generator class="uuid.hex"/>
</id>
<property name="name" index="NAME" not-null="true" type="string" column="NAME"/>
<property name="value" type="string" column="VALUE"/>
</class>
</hibernate-mapping>


Name and version of the database you are using: MsSql 2000

Hi Alex,
When you say
you mean that when A is updated, you want to avoid the B childs to be updated too. Am I right?

In that case, why would you want to do that?

Have you tried to implement the Interceptor interface? It provides callbacks so that you can explicitly control saves, updates and deletions.

Good luck!

Hi pablius,

Thanks for the reply.

You are right, I want that everything will be updated.

But the question was why there is a need in a separate UPDATE statement. The B's row can include the A's ID that can be set by a single INSERT .

Thanks,
Alex

It's because your object B doesn't contain a reference back to A.

If B had a <many-to-one> relationship pointing back to the parent, the ID of A would be assigned to B when it was generated and stored in the database. Without the relationship, there is no place to put the ID on the B object.

I'm a little confused by your mapping to be honest. Assuming B_ID is the PK of B, how is it the KEY to the B collection on A ?

Read Section 22.2 of the 3.0.5 docs for more information.

Thanks for the answer!

Sounds like I have to have use inverse attribute when declaring set of Bs:

...
<set fetch="join" lazy="false" cascade="all-delete-orphan" name="bs" inverse="true" >
<key column= "A_ID" />
<one-to-many class="myPackage.B"/>
</set>



I think with the inverse attribute there is no need in a back reference...



You are right, the mapping was incorrect... It should include A_ID as a key.



Thanks for your help,

Alex