Yeah, I often see a gap between in-memory tools like pandas and scalable processing that I thought was closed years ago, but apparently not.