I have a query similar to:

from Person p
inner join fetch p.addresses
where status='test'


I'm using the Hibernate Console plugin for eclipse, which generates the sql. I copy the sql and run it on the database and the database returns back two rows.

The only thing that is different b/w the two rows returned is address_id (primary key) and the address_type (discriminator column).

I would expect hibernate to return only one Person object, but instead I get two Person objects which are equal to each other. This seems really strange to me.

1. Am I correct in thinking I should only get back one Person object?
2. If so, then how can I go about debugging/fixing this.

thanks in advance,
Mike

You'll get duplicate results when you specify fetch on a join for a collection because the Hibernate result count always matches the JDBC row count. You can removed the duplicates by inserting them into a set.

http://www.hibernate.org/117.html#A11

ok I found a solution that is more the hibernate way of doing things. Creating a set to solve this problem just seems like it is just darn wrong.

Criteria criteria = session.createCriteria(Person.class);
results = criteria.setFetchMode("addresses", FetchMode.JOIN)
.add(Expression.eq("status", "test"))
.setResultTransformer(CriteriaSpecification.DISTIN CT_ROOT_ENTITY)
.list();

This will return only one Person object, which is exactly what I want. Now the question is...why can't you do this in HQL. Or Rather...how do you do it in HQL.

According to the following link, it looks as if hibernate 3.2 will have transformers available to the query object. Hence, the equivalent could be done with HQL.

http://blog.hibernate.org/cgi-bin/blosx ... d_sql.html

mmasters - if you look at the code for the DistinctRootEntityResultTransformer you will see that all that does is use a HashSet to create a new list internally.