Hi, >I think especially for "document like" structures, like >nt:hierarchyNode+jcr:content (eg: cq:Page+cq:PageContent) it makes >sense to store the subtree in one object, as such a document is >usually treated as 1 entity and all its subnodes are usually traversed >on read. IIRC, it was one of the initial design ideas of the MK to do >a smart subtree collating in oder to reduce the number of json objects >to be stored.
Yes. Even thought, in this case it's not about saving storage space, but using less read operations. But the idea is the same. Regards, Thomas
