I would like to run the following loop with pandas DataFrameGroupBy object within a framework of multiprocessing to reduce runtime (about 13 Mio csv-files with about 1 kb each s
