http://stackoverflow.com/questions/10604546/divide-file-up-by-values-into-equal-sections

group_them.py

import math
import itertools
import collections


def sum_column(data):
    return sum(zip(*data)[1], 0.0)


def split_groups(sensors):
    sensors.sort(key=lambda item: item[1], reverse=True)
    per_group = len(sensors) // 12
    average = sum_column(sensors) / len(sensors)
    data = collections.deque(sensors)
    groups = [[] for i in xrange(12)]
    cycle = itertools.cycle(groups)
    try:
        while True:
            current = cycle.next()
            if len(current) == per_group - 1:
                if sum_column(current) < average:
                    current.append(data.popleft())
                else:
                    current.append(data.pop())
                continue
            current.append(data.popleft())
            current.append(data.pop())
    except IndexError:
        return groups


def split_groups2(sensors):
    sensors.sort(key=lambda item: item[1], reverse=True)
    groups = [[] for i in xrange(12)]
    cycle = itertools.cycle(groups)
    per_group = int(math.ceil(len(sensors) / 3.))
    partitions = [sensors[i:i + per_group] for i in xrange(0, len(sensors)
                                                           per_group)]
    medium, low = map(reversed, partitions[1:])
    for sensor, value in itertools.chain(partitions[0], medium, low):
        cycle.next().append((sensor, value))
    return groups


def format_groups(result):
    ret = []
    for group in result:
        tmp = []
        tmp.append('\n'.join('{0}   {1}'.format(k, v) for k, v in group))
        tmp.append(' ' * 8 + str(int(sum_column(group))))
        ret.append('\n'.join(tmp))
    return '\n\n'.join(ret)


if __name__ == '__main__':
    import sys

    implementation = split_groups
    if '--second' in sys.argv:
        sys.argv.remove('--second')
        implementation = split_groups2

    with open(sys.argv[1]) as fobj:
        sensors = []
        for line in fobj:
            sensor, value = line.strip().split(',  ')
            sensors.append((sensor, int(value)))
        sys.stdout.write(format_groups(split_groups(sensors)))
        sys.stdout.write('\n')