sum_groups.py

import csv
import datetime
import time
import pprint

import itertools

def get_data():
    f = open('DATA_1334.csv')

    reader = csv.DictReader(f.read().splitlines())

    rows = []

    for row in reader:
        # extract row
        utc = row['utc']
        # parse into datetime
        format_withtime = '%m/%d/%y %H:%M'
        format_no_time = '%m/%d/%y'
        try:
            dt_struct = time.strptime(utc, format_withtime)
        except ValueError:
            dt_struct = time.strptime(utc, format_no_time)
        dt = datetime.datetime.fromtimestamp(time.mktime(dt_struct))

        rows.append({
            'datetime': dt,
            'use': row['use']
        })

    return rows

def group_by_date(rows):
    return itertools.groupby(rows, key=lambda row: row['datetime'].date())

grouped = group_by_date(get_data())

def sum_groups(groups):
    summed = []
    for date, group in groups:
        sum = 0
        for item in group:
            sum += float(item['use'])
        summed.append({
            'date': date,
            'use_sum': round(sum, 3)
        })
    return summed

sums = sum_groups(group_by_date(get_data()))

output = open('summed_1334.csv', 'w')
writer = csv.DictWriter(output, ['date', 'use_sum'])
writer.writeheader()
writer.writerows(sums)