Datetime Object Construction
From Constructor
from datetime import datetime datetime(year, month, day, hour) datetime(year, month, day, hour, minute) datetime(year, month, day, hour, minute, secons) ...
are all valid method overloading for datetime object construction.
From String
From Local Time String to Local Datetime
from datetime import datetime hk_time = "2021-11-11 10:32:34.126377" hk_time_datetime = datetime.strptime(hk_time, '%Y-%m-%d %H:%M:%S.%f')
From Local Time String to UTC Datetime
from datetime import datetime, timezone local_time = "2021-11-11 10:32:34.126377" local_time_datetime = datetime.strptime(local_time, '%Y-%m-%d %H:%M:%S.%f') # local datetime object utc_time_datetime = local_time_datetime.astimezone(timezone.utc) print(local_time_datetime.hour, utc_time_datetime.hour) # output: 10, 2
From UTC Time String to Local Datetime
from datetime import datetime, timezone from dateutil import tz utc_time = "2021-11-11 10:32:34.126377" utc_time = datetime.strptime(utc_time, '%Y-%m-%d %H:%M:%S.%f') # This is still a local datetime utc_time_datetime = utc_time.replace(tzinfo=timezone.utc) # correct time-zone info local_zone = tz.tzlocal() hk_time_datetime = utc_time_datetime.astimezone(local_zone)
From Timestamp
No matter which timezone we use, we still get the unique timestamp as below:
local = local_time_datetime.timestamp utc = utc_time_datetime.timestamp print(local, utc) # result: 1636597954.126377, 1636597954.126377
Therefore timestamp is also a good data normalization of time, and the following script will do:
local_datetime_obj = datetime.fromtimestamp(timestamp)
Also note that we have the following conversion as timestamp may be generated from frontend which use javascript:
js_timestamp = int(py_timestamp * 1000) py_timestamp = js_timestamp / 1000
Datetime Object Manipulation
Timedelta as Summand
Sometimes we want to calculate the offset time of a given time.
For example in a GPS track recording project, a trip may start at datetime start_time and end at datetime end_time. We need GNSS base stations data that are within two hours before start_time and 1 hour after end_time, then we do the following operations:
from datetime import timedelta offset_start_time = start_time - timedelta(hours=2) offset_end_time = end_time + timedelta(hours=1)
the resulting object is still a datetime object.
Now we can download the base stations data from Geodetic Survey of Hong Kong (GSHK) by converting them to utc+0 format (we can use .astimezone(timezone.utc) to generate utc datetime) and getting the corresponding hours.
Timedelta from Subtraction
Since base stations data are not immediately available at the time we stop our trip. We need to check whether the current time (the time we download the differential data) is far enough from datetime object end_time, say 15 minutes.
from datetime import datetime difference = datetime.now() - end_time # This is a Timedelta object diff_in_min = difference.total_seconds() / 60 # difference.seconds is not always what we want if diff_in_min >= 15: # download data pass