This seems to have been alluded to in [this comment|https://hibernate.atlassian.net/browse/HHH-10263?focusedCommentId=74788&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-74788] on HHH-10263:
{quote}
We discussed as a team and came to the agreement that long term the best option here is to add entity-graph support to MultiIdLoadAccess.
{quote}
I'm finding MultiIdentifierLoadAccess very useful - it would be great if at the same time as loading a bunch of `Leaf` objects by their IDs, we could initialise their `Branch` and `Tree` associations as well. I imagine usage something like this:
{code:java}
return session.byMultipleIds(Leaf.class)
.leftJoinFetch("branch as b")
.leftJoinFetch("b.tree")
.multiLoad(leafIds);
{code}
I suppose that if you joined a many-to-one association you'd have to manage the de-duplication before returning the results. |
|
