# h1. How to reproduce this issue
<strike>
[- https://github.com/cuipengfei/Spikes/tree/master/jpa/ClassIdUpdateIssue ~~ -|https://github.com/cuipengfei/Spikes/tree/master/jpa/ClassIdUpdateIssue]
- this code can reproduce the issue, just run the main method then the error will happen. </strike> -
[ https://github.com/gregturn/spring-data-jpa-id-class-issues/tree/main/src/test/java/com/example/demo |https://github.com/gregturn/spring-data-jpa-id-class-issues/tree/main/src/test/java/com/example/demo|smart-link]
Just make sure you have docker installed and run these 👆 unit tests.
The unit test with * * EmbeddedId* * will be ok.
But the one with * * IdClass* * will * * fail* * , while we * * expect* * it to be * * successful* * .
h1. Issue Description
There is a base class like this:
{code:java}
@Entity
@Inheritance(strategy = InheritanceType.SINGLE_TABLE)
@DiscriminatorFormula("case when vip_number is not null then 'vip' else 'normal' end")
@DiscriminatorValue("normal")
@IdClass(CustomerPK.class)
public class CustomerWithIdClass implements Serializable {
private String firstName;
private String lastName;
@Id
private Long versionId;
@Id
private Long unitId;
protected CustomerWithIdClass() {
}
// getter and setters ......
}{code}
Its IdClass is like this:
{code:java}@NoArgsConstructor
@EqualsAndHashCode
@Embeddable
public class CustomerPK implements Serializable {
private Long unitId;
private Long versionId;
public void setUnitId(Long unitId) {
this.unitId = unitId;
}
public void setVersionId(Long versionId) {
this.versionId = versionId;
}
}{code}
Then it has a subclass:
{code:java}@Entity
@DiscriminatorValue("vip")
public class VipCustomerWithIdClass extends CustomerWithIdClass {
private String vipNumber;
public VipCustomerWithIdClass() {
}
public VipCustomerWithIdClass(String firstName, String lastName, String vipNumber) {
super(firstName, lastName);
this.vipNumber = vipNumber;
}
public String getVipNumber() {
return vipNumber;
}
public void setVipNumber(String vipNumber) {
this.vipNumber = vipNumber;
}
}
{code}
The subclass only adds one additional field, nothing else fancy.
Then when I try to persist an instance of the subclass like this:
{code:java} CustomerWithIdClass customer = new CustomerWithIdClass("a", "b");
customer.setVersionId(123L);
customer.setUnitId(456L);
repository.save(customer);//save object of base class, ok
customer.setFirstName("a2");
repository.save(customer);//modify object of base class and save again, ok
VipCustomerWithIdClass vipCustomer = new VipCustomerWithIdClass("a", "b", "888");
vipCustomer.setVersionId(987L);
vipCustomer.setUnitId(654L);
repository.save(vipCustomer);//save object of subclass, ok
vipCustomer.setVipNumber("999");
repository.save(vipCustomer);//modify object of subclass and save again, NOT OK
// ↑ THIS FAILS BECAUSE OF PRIMARY KEY CONFLICT. INSERT STATEMENT WAS USED INSTEAD OF UPDATE, WHY?
// this failure only happens when:
// 1. base class uses IdClass for composite primary key
// 2. saving an instance of the subclass for the second time after modification
{code}
h1. The Error
Then there will be an *error of duplicate pkey* when I try to save the instance of the *subclass* for the *second time* *after some modification*.
The error seems to be related with the @*IdClass* annotation, because I have tried using @*EmbeddedId* and @*Embeddable* for composite pkey, then the error does not happen.
h1. The question
What is the root reason of this issue? Is @IdClass not supposed to be used for this scenario?
h1. links
[https://stackoverflow.com/questions/75147518/jpa-hibernate-duplicate-pkey-error-when-updating-entity-that-is-a-subclass-of|https://stackoverflow.com/questions/75147518/jpa-hibernate-duplicate-pkey-error-when-updating-entity-that-is-a-subclass-of|smart-link]
[https://github.com/spring-projects/spring-data-jpa/issues/2767|https://github.com/spring-projects/spring-data-jpa/issues/2767|smart-link] |
|
