Question

I have two entities:

Corpus entity:

@Entity(name = "CORPUS")
public class Corpus implements Serializable {

    @Id
    @GeneratedValue(strategy = GenerationType.IDENTITY)
    @Column(name = "ID")
    private Long id;

    @OneToMany(mappedBy = "corpus", cascade = CascadeType.ALL, fetch = FetchType.LAZY)
    private Collection<CorpusHistory> corpusHistories;

    //Setters and getters...
}

Corpus history entity:

@Entity(name = "CORPUS_HISTORY")
public class CorpusHistory implements Serializable {

    @Id
    @GeneratedValue(strategy = GenerationType.AUTO)
    @Column(name = "ID")
    private Long id;

    @ManyToOne(fetch = FetchType.LAZY)
    @JoinColumn(name="CORPUS_ID")
    private Corpus corpus;

    //Setters and getters...
}

The corpus entity can have many records of corpus history so I am annotating it with @OneToMany. I want the mapping to be done using the corpus id so I am using @JoinColumn(name="CORPUS_ID") and @ManyToOne annotation in corpus history entity.

Before persisting the corpus object to database I set the corpus history collection to it:

LinkedList<CorpusHistory> corpusHistories = new LinkedList<CorpusHistory>();
for (Change change : changes) {
    CorpusHistory corpusHistory = new CorpusHistory();
    //corpusHistory.setCorpusId(String.valueOf(corpusId)); ?????
    corpusHistory.setRevisionAuthor(change.getName());
    corpusHistory.setRevisionDate(change.getWhen());
    corpusHistory.setRevisionNote(change.getNote());
    //corpusHistoryFacade.create(corpusHistory);
    corpusHistories.add(corpusHistory);
}

corpus.setCorpusHistories(corpusHistories);

Records are created all it is ok but in the corpus history table the CORPUS_ID column is always null. And when I am retrieving corpus from database the history list is empty. I do not understand how can I specify the corpus id to corpus history if the corpus record is not created yet?

Isn’t this an EJB job to do? With @OneToMany and @ManyToOne mapping the appropriate ID’s should be mapped and stored into appropriate columns (in this case the corpus id should be stored in every record of corpus history column CORPUS_ID).

Or I am misunderstanding something here? I have tried many tutorials, no success… I am stuck here.

Answer 1

for (Change change : changes) {
    CorpusHistory corpusHistory = new CorpusHistory();
    corpusHistory.setCorpus(corpus);
    ...
}

The owner side is the one without the mappedBy attribute. You must initialize the owner side in order to tell JPA that the association exists. Since you have a cascade on corpus.histories, when you persist the corpus, JPA will also persist the histories.