Summing by column with pandas

month_summer.py

#!/usr/bin/env python

from collections import Counter
import pandas as pd
import glob

""" Open a bunch of .csv files which have columns that look like
    startdate, enddate, account, service, 2010-01, 2010-02, 2010-03, ...

    Sum the date-like columns across all of the .csv files and display.
"""


def main():
    bymonth = Counter()
    for csv in glob.glob('historical_payments/*.csv'):
        # skiprows is because these .csv files start with a comment line
        tab = pd.read_csv(csv, skiprows=1)
        for month, amount in tab.filter(regex='\d{4}-\d{2}').sum(axis=0).iteritems():
            bymonth[month] += amount

    for month in sorted(bymonth.iterkeys()):
        print "%s: %0.2f" % (month, bymonth[month])


if __name__ == '__main__':
    main()