(Hardy, sorry for the double post, forgot the list)
On Mon, Aug 5, 2013 at 3:20 PM, Hardy Ferentschik <ha...@hibernate.org> wrote:
> TBH i am not completely sure I understand what the method does. The docs says:
>
> "From the set of classes a new set is built containing all indexed
> subclasses, but removing then all subtypes of
> indexed entities."
>
> Not sure what exactly that means. What are you using this method for? Maybe
> you could
> provide a use case.
Well, we use it for our mass indexing magic button to index one class
at a time. Otherwise the mass indexer requires too many db connections
when we have too many entities: there are a couple of forum posts and
JIRA issues about this problem.
The purpose of the "rootification" done in the mass indexer is that if you have:
Vehicle > Car
> Truck
Animal > Cat
> Dog
the toRootEntities only returns Vehicle and Animal so that you can
send them to the mass indexer machinery. Otherwise, you will reindex
Car, Truck, Cat and Dog 2 times (one time when you deal with the
parent class and one time when you deal with the child one).
> Do you have any suggestion on how to add this into the the current API? Do
> you want to
> have something like isRoot on IndexedTypeDescriptor (thought I am not quite
> sure yet what
> "root" means in this case.
isRoot more or less means isRootInIndexedClassHierarchy.
I thought of adding a getRootIndexedTypes() as the mass indexer code
is well tested and based on this principle but I agree it's probably
nicer to put it in IndexedTypeDescriptor.
One way or another, it would be nice to be able to do that without ugly hacks.
--
Guillaume
_______________________________________________
hibernate-dev mailing list
hibernate-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hibernate-dev
