[Python] pandas datetime 타입 시간/주/일 더하기

시스템 로그를 분석할 때, 로그시간에 UTC시간을 더해줘야 할 때가 있는데, 

 

datetime의 timedelta의 메소드를 이용하여 변환할 수 있겠다.

 

 

---

기준일

 

from datetime import datetime, timedelta

.......
print(data['time'])

0        2019-08-27 00:00:00
1        2019-08-27 00:00:00
2        2019-08-27 00:00:00
3        2019-08-27 00:00:00
4        2019-08-27 00:00:00
                 ...        
373064   2019-08-27 23:59:58
373065   2019-08-27 23:59:59
373066   2019-08-27 23:59:59
373067   2019-08-27 23:59:59
373068   2019-08-27 23:59:59
Name: time, Length: 373069, dtype: datetime64[ns]

 

 

9시간(hour) 더하기(UTC 한국시간)

 

data['time'] = pd.DatetimeIndex(data['time']) + timedelta(hours=9)
print(data['time'])

0        2019-08-27 09:00:00
1        2019-08-27 09:00:00
2        2019-08-27 09:00:00
3        2019-08-27 09:00:00
4        2019-08-27 09:00:00
                 ...        
373064   2019-08-28 08:59:58
373065   2019-08-28 08:59:59
373066   2019-08-28 08:59:59
373067   2019-08-28 08:59:59
373068   2019-08-28 08:59:59
Name: time, Length: 373069, dtype: datetime64[ns]

 

 

2일(day) 더하기

 

data['time'] = pd.DatetimeIndex(data['time']) + timedelta(weeks=1)
print(data['time'])

0        2019-09-03 00:00:00
1        2019-09-03 00:00:00
2        2019-09-03 00:00:00
3        2019-09-03 00:00:00
4        2019-09-03 00:00:00
                 ...        
373064   2019-09-03 23:59:58
373065   2019-09-03 23:59:59
373066   2019-09-03 23:59:59
373067   2019-09-03 23:59:59
373068   2019-09-03 23:59:59
Name: time, Length: 373069, dtype: datetime64[ns]

 

 

1주(week) 더하기

 

data['time'] = pd.DatetimeIndex(data['time']) + timedelta(weeks=1)
print(data['time'])

0        2019-09-03 00:00:00
1        2019-09-03 00:00:00
2        2019-09-03 00:00:00
3        2019-09-03 00:00:00
4        2019-09-03 00:00:00
                 ...        
373064   2019-09-03 23:59:58
373065   2019-09-03 23:59:59
373066   2019-09-03 23:59:59
373067   2019-09-03 23:59:59
373068   2019-09-03 23:59:59
Name: time, Length: 373069, dtype: datetime64[ns]