Hi ,

I am trying to model a One To Many relationship using a Link Table (aka Join Table) -
and I am struggling with a Cascade delete that goes horribly wrong. I am using JPA with Hibernate Annotations

Entity classes are a) Project.java , b) Employee.java . A project can have many employees in it .
Hence a Project will have a List<Employee> employees

// mapping for this in Project.java is as follows

@OneToMany(fetch=FetchType.LAZY,cascade=CascadeType.ALL)
@Cascade(org.hibernate.annotations.CascadeType.DELETE_ORPHAN)
@JoinTable(name="PROJECT_EMPLOYEE_LINK",
joinColumns=@JoinColumn(name="PROJECT_ID"),
inverseJoinColumns=@JoinColumn(name="EMPLOYEE_ID")
)
public List<Employee> getEmployees()
{
return employees;
}

The link table is a PROJECT_EMPLOYEE_LINK with ONLY 2 cols in it viz PROJECT_ID, EMPLOYEE_ID

Employee.java is a normal Entity class, with Id, firstName,lastName,designation properties
public class Employee implements Serializable
{
@Id
private Long id ;
private String firstName;
private String lastName;
private String designation;
}
Note: No reference in employee class to the parent entity , i.e. Project

In my test class I have the following method :

public void testSaveEmployeeWithProject() throws Exception
{
Project project = new Project();
project.setDescription("xproject");

Employee emp1 = new Employee("employee1");
Employee emp2 = new Employee("employee2") ;
Employee emp3 = new Employee("employee3") ;

List<Employee> list = new ArrayList<Employee>();
list.add(emp1);
list.add(emp2);
list.add(emp3);
project.setEmployees(list);
// this will save it PROJECT, EMPLOYEE ( one or more rows ) , and EMPLOYEE_PROJECT_LINK ( employee and project ID )
employeeDAO.saveProjectDetails(project);
}

The save works fine ... When I run the above test , one row in PROJECT table, 3 rows in employee table , and 3 rows in the link table ( PROJECT_EMPLOYEE_LINK ) get saved.

The problem I have is when I TRY to delete a Project - What I would expected to see is that

a. The correct Project ID is deleted in the db.

b. All the corresponding rows for the particular PROJECT_ID are removed from the PROJECT_EMPLOYEE_LINK Table . How can a link between Project and Employee exist, when there is no Project itself ?

c. Employee is left untouched... As they are separte entities that exist on their own. Possibly other entities like Payroll, HR etc could be referring to the Employee . Also this is the VERY Reason , why Employee has been modelled separately without
having any FK references to other Entities.

What happens is that rows in the Project, Employee, as well as PROJECT_EMPLOYEE_LINK all get deleted .

Adding the @Cascade(org.hibernate.annotations.CascadeType.DELETE_ORPHAN) ensures that the
rows in the PROJECT_EMPLOYEE_LINK table get deleted . This is bcos I read somewhere that JPA on its own does nt support the equivalent of the Hibernate all delete orphan , and hence this custom Hibernate annotation is required.

Table structure is as follows

EMPLOYEE
---------
ID
FIRST_NAME
LAST_NAME
DESIGNATION

PROJECT
---------
ID
PROJECT_NAME
DESCRIPTION

PROJECT_EMPLOYEE_LINK
--------------------------
PROJECT_ID references PROJECT.ID
EMPLOYEE_ID references EMPLOYEE.ID

I am using JPA 1 , Hibernate 3.2.3.ga,Hibernate Annotations 3.3.1.ga,Hibernate Entity Manager 3.3.2.ga, javax-persistance 1.0, Spring 2.5.6

Any help will be greatly appreciated

Many thanks

Kasi

Replying to my own Question posed above -

The Solution can be obtained, If I use a Hibernate Specific annotation (ie org.hibernate.annotations.CascadeType.SAVE_UPDATE ) and not a JPA (javax.persistence.Cascade.* ) one when I specify the cascade attribute in the One To Many side of the relationship

The Code would hence be as follows in Project.java

@OneToMany
@Cascade(org.hibernate.annotations.CascadeType.SAVE_UPDATE)
@JoinTable(name="PROJECT_EMPLOYEE_LINK",
joinColumns=@JoinColumn(name="PROJECT_ID"),
inverseJoinColumns=@JoinColumn(name="EMPLOYEE_ID")
)
public List<Employee> getEmployees() {
return employees;
}

This would entail , the following

1. Project is deleted.
2. Corresponding rows in the Link Table are also deleted
3. Employee rows are left untouched .

Hope this helps someone who is modelling a One-To-Many using a Link Table

Thanks - Kasi