纪元时间和正常日期之间的差异（以秒为单位）-python

Question

I have two variables : One is epoch time and other is normal datetime,

df['epochtime']

0       1457586382
1       1457586382
2       1457586391
3       1457586692
4       1457586391
5       1457586391
6       1457586692
7       1457586692
8       1457572611

df['time']
0      2016-03-10 00:06:21.903
1      2016-03-10 00:06:21.908
2      2016-03-10 00:06:30.895
3      2016-03-10 00:06:30.895
4      2016-03-10 00:06:30.899
5      2016-03-10 00:06:30.899
6      2016-03-10 00:06:31.045
7      2016-03-10 00:06:31.094
8      2016-03-10 01:16:51.390 

I want the difference between these two times in seconds. The following is my trying,

pd.to_datetime(df['epochtime'], unit ='s' ) - df['time']

The following is my output,

0               00:00:00.097000
1               00:00:00.092000
2               00:00:00.105000
3               00:05:01.105000
4               00:00:00.101000
5               00:00:00.101000
6               00:05:00.955000
7               00:05:00.906000
8      -1 days +23:59:59.610000   

I want to make this as integer in seconds. Also for the 8th value for some strange reason getting difference as -1 days, where it should be seconds only. Can anybody help me getting the seconds difference as integer ?

Answer 1

You have to convert datetime to epoch time, and then do the arithmetic subtraction.

As you said you are always use UTC time, you need to use calendar.timegm() because it takes UTC time as parameter.

import time
import datetime
import calendar

TIME_FORMAT = '%Y-%m-%d %H:%M:%S'

def local_datetime_2_timestamp(local_datetime):
    '''strptime takes local datetime as parameter'''
    timestamp = str(time.mktime(datetime.datetime.
        strptime(local_datetime, TIME_FORMAT).timetuple()))[:-2]
    return timestamp

def utc_2_timestamp(utc_datetime):
    '''timegm takes UTC time as parameter'''
    return calendar.timegm(utc_datetime.timetuple())

print local_datetime_2_timestamp("2016-03-10 00:06:21.903".split(".")[0])
print utc_2_timestamp(datetime.datetime(2016,3,10,0,6,21))

Answer 2

from time import mktime, strptime

date_a = '20-01-2015 11:05:02'
epoch_b = 1457586692

epoch_a = int(mktime(strptime(date_a, '%d-%m-%Y %H:%M:%S')))

print(epoch_a - epoch_b)