From a high level my code:

1. Runs an aggregation using "aggregate_raw_batches"
2. Maps the resulting cursor to a RawBSONDocument
3. Run an insert_many using the results of #2
4. Checks the results of the insert_many for the "inserted_id" field.
5. Find that this field is always an empty array even when I verified it is correctly inserting documents.

raw_cursor = my_coll.aggregate_raw_batches(multi_sample_pipeline,allowDiskUse=True,maxTimeMS=180)
mapped_docs = map(RawBSONDocument,pedigree_results_bson_cursor)
results = diff_coll.insert_many(mapped_docs,ordered=False)

print(len(results.inserted_ids)) # Seemingly always 0, even when inserting non zero number of documents.