This feature would enable MongoDB to automatically generate common patterns by computing entire collections. A particularly useful application would be the ability to identify duplicate fields within a collection and create a computed collection that groups all possible duplicate together.

Sometimes, a field is duplicated between two documents, while another field in the same document is duplicated across seven other documents. Each of these seven documents might have fields that are duplicated in yet other documents, leading to a vast iteration in search of possibilities.

Having MongoDB abstract all this logic in a performant manner to create computed collections would be incredible! It would significantly ease the developer's workload.