Hey.. i have a silly question.. i have a huge csv file that i read into memory using fread upfrontly allocating memory with malloc and then i store each column as a separate file .. after that i parallely invoke processes each calling a python script that reads two of the columns and calculates some statistics using group by from pandas.. while that is fairly ok, i am looking for improvement.. thus i am wondering - how could i skip the write column part and access malloced object directly from ram in python?