集合未充满OneToMany关系-EclipseLink

Question

I'm mapping a database (Oracle 11g) to JPA entities using EclipseLink. I have mapped almost every table to JPA objects but I've just found a problem:

In the \\Curve\\ entity I have the following fields:

@Id
@Column(name = "COD_CURVE")
@GeneratedValue(strategy = GenerationType.SEQUENCE, generator = "seq.gen")
@SequenceGenerator(name = "seq.gen.curve", sequenceName = "SEQCURVE", allocationSize = 1)
private long codCurve;

@Id
@Column(name = "FEC_HISTORIC")
@Temporal(javax.persistence.TemporalType.DATE)
private Date fecHistoric;

@OneToMany(mappedBy="codCurve", cascade = CascadeType.ALL, fetch = FetchType.LAZY)
private Collection<CurveDetail> detailsCollection;

In the \\CurveDetail\\ entity I have this:

@ManyToOne
@JoinColumns({
    @JoinColumn(name = "COD_CURVE", referencedColumnName = "COD_CURVE"),
    @JoinColumn(name = "FEC_HISTORIC", referencedColumnName = "FEC_HISTORIC")
})
private Curve codCurve;

The problem is that when I query the \\Curve\\ entity, the details always are null, despite the fact that there is valid data in both tables. Checking the database I've noticed that there are no foreign key constraints in the \\CurveDetail\\ table, so I wonder ¿Are these constraints required to map the database correctly? I haven't tried to add the FK constraint myself because I'm not allowed to (have to ask a DBA to do it, and it'll take a week).

Thanks in advance!

Answer 1

Having a foreign key is not required.

Check the SQL that is generated by enabling logging ("eclipselink.logging.level"="finest")

Try executing the same SQL with the same database to see if the data exists.

Also ensure you are not corrupting the shared cache by inserting/updating an object with a null collection. You must maintain both sides of a bi-directional relationship. You could try disabling the cache to see if this is what you are doing.

Answer 2

As I know having a FK constraint is not mandatory. Once I had the same problem (but it was Oracle 9i) and found out that the name of Entity class should match the name of the table you want to map, and also JPA SQL and Hibernate HQL are both case sensitive so be careful in writing @Column tag.