read_csv_describe.py

# Reading multiple .csv files
# files/input.cova
# files/input.covb
# .....

import dask.dataframe as dd

df = dd.read_csv('files/input.cov*', names=['chr','pos','cov'], sep='\t')

print("Descriptors: {0}".format(df['cov'].describe().compute()))

//
# Operation on a single .csv that will be partioned by Dask:
import dask.dataframe as dd

df = dd.read_csv('files/out.cov', names=['chr','pos','cov'], sep='\t', blocksize=34000000) # blocksize controls the size of each partition.

print("Descriptors: {0}".format(df['cov'].describe().compute()))