I understand that estimatedDocumentCount() uses metadata to count, which makes it faster when there are a lot of documents. But the drawback is that you cant ad
estimatedDocumentCount()