Hibernate Issue : Foreign key must have same number of columns as referenced primary key

The error speaks for itself, in order to reference a composite primary key, you need a composite foreign key. (The composite primary key states that you need an unique combination of 2 fields to make a key – then you can’t possibly reference a unique key with just 1 column.)

As to how this is achieved by using xml mapping files i’m not sure, most people use annotations these days..

As for your Java classes, i’m assuming that ImportJobManagement holds an ImportJob, so then the class should not reference the id, but the object itself, like this:

public class ImportJobManagment implements Serializable {
    private ImportJob importJob;
    ...
}

The Java classes should just reference the other class, and not the members of the composite key – it is up to the mappings to map from the composite key to the Java member variable.

Answer to update:

The short answer is no, you can’t. The way a foreign key works, it lets you reference a specific row in a table. And, in order to be sure to reference a specific row, you need an identity, something that would describe only one row, and no other. In SQL there is a construct for achieving this, namely the unique key. By stating that a column (or a composite of columns) is unique, you know that its/their combined value will be unique, there would be a maximum of 1 row in the entire table with this value, anything else would be a constraint violation.

So, a foreign key refers to either a single unique-constrained column, or a composite unique-key spanning multiple columns. Since all tables has a unique key already, the primary key (which is always unique), it is common to use this for foreign key references, but any unqiue column would work.

The easiest case is when we want to reference a table with a single-column unique key. Two simple tables A, which holds a single column ‘id’, and B, which holds an ‘id’ column, but also another column, a_id, which has a foreign key to the ‘id’ column of A. An instance of this situation could be this:

A:
| id | 
|----|
|  1 |
|  2 |
|  3 |

B:
| id | a_id |
| 2  |  3   |
| 3  |  1   |

Here, each row in B reference a row in A. Its a direct reference, the value in a_id in table B corresponds directly to the value in A’ ‘id’ column. So the B with id 2 reference the A with id 3, and so on.

Now lets look at how to reference a table with a composite unique key. Lets keep our example, but now A has another column ‘sec_id’, which together with ‘id’ compose a composite primary key.

A:
| id | sec_id |
|----|--------|
| 1  |   3    |
| 3  |   1    |
| 3  |   7    |

B:
| id | a_id |
|----|------|
| 2  |  3   |

In this situation, we have a problem in B. Since a foreign key must reference a single row in the table that its referencing, this clearly does not work. Which row in A does the value ‘3’ represent? The sec_id in the first row? The id in the second or third (but in that case, which one?)? The answer is of course neither, there is not enough information in B to reference a single row in A, so SQL just won’t have it. Adding such a foreign key is thus not allowed.

In order for B to reference A, it will require both a reference to A’s ‘id’-column, and A’s ‘sec_id’ column, because a single row in A is identified by its unique combination of (‘id’, ‘sec_id’) pairs. So with B looking like this:

| id | a_id | a_sec_id |
|----|------|----------|
| 1  |  1   |     3    |
| 2  |  3   |     1    |
| 3  |  3   |     7    |

Now, B holds enough information to reference a single row in A, and as the data shows, it does.

Update again:

I’m currently reading for JPA certification, and have reached the chapter on composite key mappings. In order to map a composite primary key, you need a primary key class which maps the attributes of your key. There are two ways of doing it, one where the key attribute must also be mapped in the entity itself, and one where it’s used as an embedded key.

I’ll provide the code examples, they pretty speak for themselves (it’s using annotations, you really should do that as well.)

The first example is the base example, with a regular id-class (not embedded). Here, we have an Employee entity where the primary key is composed of an integer id and a country (two Employees can have the same id if in different countries).

@Entity
@IdClass(EmployeeId.class)
public class Employee {
    @Id private String country
    @Id
    @Column(name = "EMP_ID")
    private int id;
    private String name;
    ...
}

public class EmployeeId implements Serializable {
    private String country;
    private int id;

    public EmployeeId() { }
    public EmployeeId(final String country, final int id) {
        this.country = country;
        this.id = id;
    }

    //getters for the properties

    public boolean equals(final Object other) {
    //must be implemented
    }

    public int hashCode() {
    //must be implemented
    }
}

Note:

• The @IdClass-annotation on the class.
• Both the attributes of the id-class must also be defined in the entity
• The id-class must implement equals and hashcode

The other way this can be done is through an embedded id-class:

@Entity
public class Employee {
    @EmbeddedId
    private EmployeeId id;
    private String name;

    public Employee(final String country, final int id) {
        this.id = new EmployeeId(country, id);
    }

    public String getCountry() {
        return id.getCountry();
    }
}   

@Embeddable
public class EmployeeId {
   private String country;
   @Column(name = "EMP_ID")
   private int id;

   //constructor + getters + equals +hashCode
}

Note:

• No need to define the id-attributes in the entity
• To get attributes of the id-class from the entity, you need to get them from id-class

I like the latter better because its more compact and doesn’t contain duplication, but then again I don’t know how using the two compares..