def convert_file(fname, out_arr):
def get_date(item):
dt = None
try:
dt = datetime.strptime(item, "%m/%d/%y %H:%M")
except:
try:
dt = datetime.strptime(item, "%y-%m-%d %H:%M")
except:
try:
dt = datetime.strptime(item, "%Y-%m-%d %H:%M")
except:
dt = datetime.strptime(item, "%m/%d/%Y %H:%M")
return dt
i = 0
f = open(fname, 'rU')
reader = read_csv(f, engine='c', header=0)
for _, row in reader.iterrows():
r = BikeRide()
tm = row['Duration']
tm = tm.replace('h ', ':').replace('m ',
':').replace('sec.',
'').replace('s', '')
h, m, s = map(int, tm.split(':'))
r.duration = h * 60 * 60 + m * 60 + s
r.start_date = get_date(row['Start date'])
r.end_date = get_date(row['End date'])
try:
r.start_station = find_station(row['Start Station'])
r.end_station = find_station(row['End Station'])
except:
print 'couldnt find', row['Start Station'], 'or', row[
'End Station']
continue
r.subscribed = False if row['Subscription Type'] == 'Casual' else True
out_arr.append(r)
i += 1
if i % 10000 == 0:
print i
f.close()
dct = ijson.items(f, 'item')
# ijson lets us stream the file rather than open it all at once --- too big
i = 0
failed = 0
starting = 2453922
for j in range(starting):
i += 1
dct.next()
for r in dct:
i += 1
if i % 10000 == 0:
print i
ride = BikeRide()
st_dt = datetime.strptime(r['start date'], "%Y-%m-%dT%H:%M:%S")
ed_dt = datetime.strptime(r['end date'], "%Y-%m-%dT%H:%M:%S")
ride.start_date = st_dt
ride.end_date = ed_dt
ride.duration = r['duration']
ride.subscribed = False if r['user type'] == 'casual' else True
try:
ride.start_station = find_station(r['start station'])
ride.end_station = find_station(r['end station'])
except:
failed += 1
print 'fail number', failed
failed_ls.append(r)
continue
ride_ls.append(ride)
