csv_chunks.py

import pandas as pd
from sys import argv

def main(INFILE, CHUNK_SIZE):
    df = pd.read_csv(INFILE)

    # Assign "groups" to each row
    def assign_chunk(x):
        x += 1
        mod = x % (-1 * CHUNK_SIZE)
        diff = x - mod
        scaled = diff / CHUNK_SIZE
        return int(scaled)

    chunks = [assign_chunk(x) for x in df.index.values]
    df['chunk'] = pd.Series(chunks)
    n = chunks[-1] # This gets the last value, which will be the max

    for i in range(1, n + 1):
        tmp = df.loc[df['chunk'] == i].copy()
        OUTFILE = INFILE.replace('.csv', ' Chunk {}.csv'.format(i))
        tmp.to_csv(OUTFILE)

if __name__ == '__main__':
    INFILE = argv[1]
    if len(argv) > 2:
        CHUNK_SIZE = int(argv[2])
    else:
        CHUNK_SIZE = 100

    main(INFILE, CHUNK_SIZE)