When doing a query like this (using Nhibernate 2.1.2):
ICriteria criteria = session.CreateCriteria<MyRootType>()
.SetFetchMode("ChildCollection1", FetchMode.Eager)
.SetFetchMode("ChildCollection2", FetchMode.Eager)
.Add(Restrictions.IdEq(id));
I am getting multiple duplicate objects in some cartesian fashion. E.g. if ChildCollection1
has 3 elements, and ChildColection2
has 2 elements then I get results with each element in ChildColection1
one duplicated, and each element in ChildColection2
triplicated! This was a bit of a WTF moment for me...
So how to do this correctly?
Update:
As per Felice's suggestion, I tried using the DistinctRootEntity transformer, but this is still returning duplicates. Code:
ICriteria criteria = session.CreateCriteria<MyRootType>()
.SetFetchMode("ChildCollection1", FetchMode.Eager)
.SetFetchMode("ChildCollection2", FetchMode.Eager)
.Add(Restrictions.IdEq(id));
criteria.SetResultTransformer(Transformers.DistinctRootEntity);
return criteria.UniqueResult<MyRootType>();
You can select single distinct results by specifying SetResultsTransformer(Transformers.DistinctRootEntity);
You're doing a cartesian product here. Don't. Instead, fetch each collection separately.
BTW: this isn't something NHibernate-specific, the same applies to any ORM in any platform, or even pure SQL without any ORM at all. In general, you don't want to fetch N*M rows when you can fetch N+M instead.
I am not using NHibernate (rather, regular 'ol Hibernate), but in the Java version of Hibernate you can define One-To-Many collections as either Lists or Sets. If you define them as Sets you will not get duplicates. Surprisingly enough, this happens without overriding equals(). Unfortunately, I have the same problem and I want Lists so I can integrate with Wicket but no duplicates...
What were you mapping your collection as? If your using a bag for example you will get duplicates. You can use a set instead and you won't get duplicates.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With