I guess I’d ask
1.) Under what circumstances you’d be modifying the contents of the join table directly? By its nature, a join table is comprised of a compound primary key, and little else. It’s generally a bad idea to create a data model in which you’re modifying primary keys… so if you’re doing this often, you might want to rethink your data model a bit.
2.) Assuming that you’re not directly modifying the contents of the join table, then there should be little need for a ModifiedBy column. That leaves CreatedBy. And I can’t think of an easy solution for that.

If it’s critical to have an audit trail of all inserts and updates, you might consider storing all of the changes in a separate auditing/log table. You could then programatically populate the audit table using event listeners and/or specialized code. E.g,. if a user were to add a Topping to a Pizza, you could capture that event in a listener and write a row to an audit table with the username of the user who performed that action.

hth!

Caused by: org.apache.openjpa.persistence.PersistenceException: ORA-01400: cannot insert NULL into (“FSTCTLADM”.”PIZZA_TOPPING”.”PIZZA_TOPPING_ID”)
{prepstmnt 1967682888 INSERT INTO PIZZA_TOPPING (pizza_id, topping_id) VALUES (?, ?) [params=(int) 7, (int) 1]} [code=1400, state=23000]

The thing is, in Oracle we need to use sequence to generate the id. But I couldn't define the @manyToMany annotation to use
the sequence. I don't know how to do it.
Please let me know if there is a way in JPA to define the ID generator for the join table used in
ManyToMany cases.

To overcome this, I wrote a trigger for the join table PIZZA_TOPPING, to generate the ids using the sequence.
Then the application worked.

Still, I have a question:
Let us consider I am adding 'CreatedBy' and 'ModifiedBy' columns in all the tables to be
used for audit purposes. Ideally these columns would be populated by the user id who has logged into the application.
This will work for all the tables except the join table PIZZA_TOPPING.
Because, the application is going to supply only the
'pizza_id' and the 'topping_id' defined in the @JoinTable annotation.
And my trigger is going to supply only the 'pizza_topping_id'
using a sequence. But how do the CreatedBy and ModifiedBy columns will be populated? The trigger can not do that as it wouldn't be aware of the fact that which user has logged into the application. Is there a way to populate additional fields in a join table used in manytomany case with JPA??

Plz let me know your thoughts.

Cheers
Madhan

@Id
@GeneratedValue(generator=”SEQUENCE_NAME”)
@SequenceGenerator(name=”SEQUENCE_NAME”,sequenceName=”SEQUENCE_NAME”)

My question is, how do I generate the ids for the mapping table (PIZZA_TOPPING)?
Thanks in advance. Looking forword for your other posts.

Cheers
Madhan

Thanks for good blog entry