[jira] Created: (OPENJPA-1645) Instance not found when using inheritance and hierarchy is higher than 2

["Georg Nozicka (JIRA)" <ji...@apache.org>]

Instance not found when using inheritance and hierarchy is higher than 2
------------------------------------------------------------------------

                 Key: OPENJPA-1645
                 URL: https://issues.apache.org/jira/browse/OPENJPA-1645
             Project: OpenJPA
          Issue Type: Bug
    Affects Versions: 2.0.0
         Environment: Operating System:
Ubuntu 
Release 9.10
Kernel Linux 2.6.31-21-generic
GNOME 2.28.1

Java:
JVM in use = Sun JVM 1.6.0_20

Server:
Apache Geronimo V2.2

Open JPA:
openjpa.Runtime - Starting OpenJPA 2.0.0-SNAPSHOT

DB & JDBC:
databaseProductName: PostgreSQL
databaseProductVersion: 8.4.3
driverName: PostgreSQL Native Driver
driverVersion: PostgreSQL 8.2 JDBC3 with SSL (build 508)
            Reporter: Georg Nozicka


Assume the following class hierarchy: A super class of B superclass of C using single table strategy.

Case 1:
1a: Create and persist instance of B (e.g. primary key = 1)
1b: Find instance via entity manager find like em.find(B.class, 1) → this works as expected
1c: Find instance via entity manager find like em.find(A.class, 1) → this works as expected

Case 2:
2a: Create and persist instance of C (e.g. primary key = 2)
2b: Find instance via entity manager find like em.find(C.class, 2) → this works as expected
2c: Find instance via entity manager find like em.find(A.class, 2) → this works as expected
2d: Find instance via entity manager find like em.find(B.class, 2) → instance is not found

The reason is as follows (only part of the code is shown to make this shorter)

// TestA is like
@Entity
@Inheritance
@DiscriminatorColumn(name = "dis", discriminatorType = DiscriminatorType.STRING)
public abstract class TestA implements Serializable {

	private static final long serialVersionUID = -4733743965438652043L;

	// id
	@Id
	private long id;

	private String name;

	public TestA() {
	}
		
	public TestA(String name) {
		id = System.nanoTime();
		this.name = name;
	}

// TestsB is like
@Entity
public  class TestB extends TestA {

	private static final long serialVersionUID = -2091733862251849317L;

	// Constructors
	public TestB() {
		super();
	}

	public TestB(String name) {
		super(name);		
	}

// TestC is like
@Entity
public  class TestC extends TestB {

	private static final long serialVersionUID = -2091733862251849317L;	

	// Constructors
	public TestC() {
		super();
	}

	public TestC(String name) {
		super(name);		
	}	


The facade looks like this:
@Override
public void create(TestA test) {
	em.persist(test);
}

@Override
public <T extends TestA> T find(Class<T> cls, long id) {
	return em.find(cls, id);
}

// Somewhere in a bean ...
@EJB
private TestFacadeLocal tf;

// ...

// *** Case 1 ***

// Case 1a
TestB tb0 = new TestB("name1");
tf.create(tb0);

// Case 1b
tb0 = tf.find(TestB.class, tb0.getId());

Resulting SQL:
SELECT t0.dis, t0.name FROM TestA t0 WHERE t0.dis = ? AND t0.id = ? [params=(String) TestB, (long) 4013390270882]

The discriminator column is taken into account, this is ok

// Case 1c
TestA ta0 = tf.find(TestA.class, tb0.getId());

Resulting SQL:
SELECT t0.dis, t0.name FROM TestA t0 WHERE t0.id = ? [params=(long) 4013390270882]

The discriminator column is not taken into account to be able to retrieve TestA. This is ok.

// *** Case 2 ***

// Case 2a
TestC tc1 = new TestC("name1");
tf.create(tc1);


// Case 2b
tc1 = tf.find(TestC.class, tc1.getId());

Resulting SQL:
SELECT t0.dis, t0.name FROM TestA t0 WHERE t0.dis = ? AND t0.id = ? [params=(String) TestC, (long) 4013797794600]

The discriminator column is taken into account. This is ok

// Case 2c
TestA ta1 = tf.find(TestA.class, tc1.getId());

Resulting SQL:
SELECT t0.dis, t0.name FROM TestA t0 WHERE t0.id = ? [params=(long) 4013797794600]

The discriminator column is not taken into account to be able to retrieve TestA. This is ok.

// Case 2d
TestB tb1 = tf.find(TestB.class, tc1.getId());

Resulting SQL:
SELECT t0.dis, t0.name FROM TestA t0 WHERE t0.dis = ? AND t0.id = ? [params=(String) TestB, (long) 4013797794600]

Here the problem arises. The discriminator column is taken in the wrong way into account. It just checks for "TestB" but it should also take "TestC" into account. The result is, that the instance is not found. Actually the same scenario works fine with SNAPSHOT 1.3.0. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.