Currently, the py pins driver reads index 0 as the column, and driver save keeps the index (pandas saves index as an unnamed first column, which I think some R csv saving functions do also?). Let's ensure roundtrip by dropping the index on save, etc...
