convert_time.py

#!/usr/bin/python

import os
import csv
import math
import datetime
from optparse import OptionParser

TIME_FORMAT = '%Y,%m,%d,%H,%M,%S'
TIME_HEADERS = 'year,month,day,hour,minute,second'

parser = OptionParser()
parser.add_option("-d", "--dispose",
                  action="store_true", dest="dispose_utc", default=False,
                  help="will dispose of the UTC column");
parser.add_option("-f", "--file",
                  action="store", type="string", dest="fileName", default="results.csv",
                  help="specify the csv file you would like to process");                  
      
(options, args) = parser.parse_args()

indexOfUTC = -1; # assume non-existant
rowsOfFile = []  # start empty
fileName = options.fileName # simply the file name to be read and written to

if(os.path.exists(fileName)):
	with open(fileName, 'rb') as csvfile:
		spamreader = csv.reader(csvfile, delimiter=' ', quotechar='|')
		i = 0
		for row in spamreader:
			row_to_add = []
			
			if len(row) <= 0 or not ',' in row[0]:
				continue
				
			array = row[0].split(',')
			for val in array:
					val.strip()
					row_to_add.append(val)
					
			if i == 0:
				# Headers
				
				if "created_utc" in array:
					indexOfUTC = array.index("created_utc")
				else:
					raise RuntimeError('Invalid headers: \'created_utc\' column not found!')
				
				for header in TIME_HEADERS.split(','):
					header.strip()
					row_to_add.append(header)
					
			else:
				# Regular data lines
				
				utc = datetime.datetime.utcfromtimestamp(int(math.floor(float(array[indexOfUTC]))))
				time_data = utc.strftime(TIME_FORMAT)
				
				for val in time_data.split(','):
					val.strip()
					row_to_add.append(val)
					
			
			rowsOfFile.append(row_to_add)	
			i = i + 1
			
	csvfile = open(fileName.replace('.csv', '_converted.csv'), 'wb')
	spamwriter = csv.writer(csvfile, delimiter=',',quotechar='|', quoting=csv.QUOTE_MINIMAL)
	
	for row in rowsOfFile:
		spamwriter.writerow(row)

	csvfile.close()