Newbie.. Confused with hibernate cascade and inverse.

Ravi9 · **Joined:** Wed Oct 08, 2008 1:54 am **Posts:** 2

Hi All,

I always wondered what cascade and inverse meant. I will tell what I have understood. Can someone please tell me if I am right or wrong?

Assumption: Say there are two tables, parent and child. Parent has cid foreign key to child table's primary key childid.

Cascade: None--> If Parent table is inserted with values, then child table is not created if it does not exist and is not inserted with any value. cid is not propagated to childid

Cascade: All--> If Parent table is inserted with values, then even tho child is not already existing, it is created and child table's childid is inserted with cid of the Parent table

Inverse:true--> Both tables take responsibilities? But for what?
Inverse:false--> Only one table takes responsibility , Which one? and for what?

Thanks.

gonzao_diaz · **Joined:** Mon Nov 26, 2007 2:29 pm **Posts:** 443

"cascade" on a collection of children, means that, whatever operation you perform in the parent, it will pass on to the children.

This is unrelated to DDL. Table creation, etc, is managed by other settings.

Id's are also unrelated to this command, they can be assigned or auto-generated, according to the strategy you specify in the mapping file.

cascade=none means that no parent operation is repeated on the children.
cascade=all means that all operations (save, persist, update, delete, merge) are repeated on the children.

The "inverse" property is used in bi-directional associations. This is, relations for which you have both parent.getChildren() and child.getParent().
When you are adding a new child, Hibernate expects you to take care, from the Java point of view, of both aspects of this association.
In other words, you have to do not only this

Code:

1. parent.children.add(child)

but also this

Code:

2. child.setParent(parent)

Hibernate has to detect that a child has been added, and generate the proper SQL. The problem is when to generate that SQL. In a unidirectional relationship, the only code you need is 1, so the SQL is generated there.
In a bidirectional relationship, you have both code 1. and 2., and Hibernate has to choose at which moment it generates the SQL.
If we set 1. (the one-to-many side) as "inverse", Hibernate will only generate SQL code when 2. occurs.

Ravi9 · **Joined:** Wed Oct 08, 2008 1:54 am **Posts:** 2

Hi,

Thanks for the reply. Can you give me another example maybe for inverse? I did not get it. I understood cascade tho.

Thanks.