Is your feature request related to a problem? Please describe.

I want to test the performance of different encoders. To this end, I would like to create a collection containing payloads of a number of documents and then create different collections holding vectors created by different encoders. My life would be easier if there was a way to iterate all available records.

Describe the solution you'd like

Something like collection.keys()

I'm not too familiar with RocksDB, but I know that it provides a way to get all keys, see for example https://python-rocksdb.readthedocs.io/en/latest/tutorial/#iteration