Hello everybody!
I have simple one to many relationship:

@Entity
class Parent {
@Id
@GeneratedValue
int id;

@OneToMany
@IndexedEmbedded
Set<Child> children;
String name;
}

@Entity
class Child{
@Id
@GeneratedValue
int id;
String name;
}

I can create index, everything looks good. But I just realized that Child.Id is stored in Parent document.
The number of children could be big, so my document is getting big also. I just want to index Parent by Child name, I don't want to waste space for Child ids. I don't need them. So, I have two questions
1. Why Child id is stored in Parent document?
2. Is there a way to remove Child id from Parent document?

Thank you,
Andrey

There are several reasons, but I'm not sure about if they are still all applying to latest versions; I'll raise the issue on the developer mailing list. Feel free to join ;)
http://www.hibernate.org/community/mailinglists


There's no option to control that with the current implementation of @IndexedEmbedded, you could of course use a custom fieldbridge instead of that and then add all of child's names. Simple workaround, but viable only if you don't have many fields to add as it would quickly become complex code (you have only "name" right?).

Is this really significant, what's the impact on your index size?

Thank you very much. Things are actually worse then I described. My data is highly normalized and I have to get the "name" field from a grandchild. Which means I have two useless ids per object. Total could be up to several hundred. I will look into FieldBridge, it seems to be what I need.
Thank you,
Andrey