On document store indexes the article says: > If all order data is lumped together, the user will be forced to retrieve the entirety of each order that contain

On document store indexes the article says: > If all order data is lumped together, the user will be forced to retrieve the entirety of each order that contains a relevant order line - or perhaps even to scan the entire database and examine every order to see whether it contains a relevant order line. Both of these are untrue. Author needs to read up on secondary indexes.