Today when a list of documents is requested, first a request for their ids is made, then N requests for each of their schema's tables is done (N+1 problem).
This optimization will get all properties in a given schema for all the documents with only one request.