Handling date time in code becomes little tricky when the project is used internationally because then there comes additional term Timezone. Timezone is a property of a location which needs to be considered while comparing that time with the time in another location. For example – there are two villages A and B. One day Ram from village A calls his friend Shyam in village B at 8:00 am to wish “good morning”. But Shyam receives Ram’s call at 6pm on same day and he replies “good evening”. That means village A’s timezone is 10 hrs behind village B’s timezone. So here we need some reference timezone about which all other timezones can be declared. This is where UTC (Coordinated Universal Time) comes into play. UTC is reference timezone by which all the timezones are declared. For example – Indian timezone is 5 hrs and 30 mins ahead of UTC which is denoted as UTC+05:30. In languages, these timezones are declared in date time library using constants such as ‘Asia/Kolkata’ which is Indian Standard Time. I will be talking about working with date time in python in this blog. In the FOSSASIA’s Open Event project since it is event management system, handling date-time with the timezone is one of the important tasks.
Here is the relevant code:
>>> import datetime >>> import pytz >>> now = datetime.datetime.now()
datetime.datetime.now() returns naive datetime as of the time setting of the machine on which the code is running. Naive date means it doesn’t contain any info about the timezone. It just contains some number values of year, month, hours etc. So just by looking at naive date we cannot actually understand the time. There comes aware datetime which contains timezone info.
>> now datetime.datetime(2017, 5, 12, 21, 46, 16, 909983) >>> now.isoformat() '2017-05-12T21:46:16.909983' >>> aware_now = pytz.timezone('Asia/Kolkata').localize(now) >>> aware_now datetime.datetime(2017, 5, 12, 21, 46, 16, 909983, tzinfo=<DstTzInfo 'Asia/Kolkata' IST+5:30:00 STD>
Pytz provides timezone object which takes string argument for timezone which has localize method which adds timezone info to the datetime object. Hence now aware datetime has timezone info too. Now if we print the time.
>>> aware_now.isoformat() '2017-05-12T21:46:16.909983+05:30
We get the +05:30 extra string at the end which gives timezone info. +05:30 means the timezone is 5 hrs and 30 mins ahead of UTC timezone. The comparison between datetimes can be made between naive-naive and aware-aware. If we try to compare between naive and aware,
>>> now < aware_now Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: can't compare offset-naive and offset-aware datetimes >>> now2 = datetime.datetime.now() >>> now2 datetime.datetime(2017, 5, 15, 9, 44, 25, 990666) >>> now < now2 True >>> aware_now.tzinfo <DstTzInfo 'Asia/Kolkata' IST+5:30:00 STD>
tzinfo carries timezone info of the datetime object. We can make aware date to unaware by method replacing tzinfo to None.
>>> unaware_now = aware_now.replace(tzinfo=None) >>> unaware_now.isoformat() '2017-05-12T21:46:16.909983'
Formating datetime is done by mostly these two methods. One of which takes string format in which the result is required and another returns datetime in iso format.
>>> now.strftime('%Y-%m-%dT%H:%M:%S%z') '2017-05-12T21:46:16' >>> aware_now.strftime('%Y-%m-%dT%H:%M:%S%z') '2017-05-12T21:46:16+0530' >>> now.time().isoformat() '21:46:16.909983' >>> now.date().isoformat() '2017-05-12' >>> now_in_brazil_east = datetime.datetime.now(pytz.timezone('Brazil/East')) >>> now_in_brazil_east.isoformat() '2017-05-15T06:49:51.311012-03:00'
We can pass timezone argument to the now method to get current time in the passed timezone. But the care must be taken as this will use the timezone setting of the machine on which code is running to calculate the time at the supplied timezone.
In the FOSSASIA’s Open Event Project, date-time is taken from the user along with the timezone like in one of the example shown below.
This is part of the event creation page where user has to provide date-time along with the timezone choice. At back-end the date-time and timezone are stored separately in the database. The event model looks like
class Event(db.Model): """Event object table""" ... start_time = db.Column(db.DateTime, nullable=False) end_time = db.Column(db.DateTime, nullable=False) timezone = db.Column(db.String, nullable=False, default="UTC") ...
The comparison between the stored date-time info with the real time cannot be done directly. Since the timezones of the both times need to be considered along with the date-time values while comparison. Likewise there is one case in the project code where tickets are filtered based on the start time.
def get_sales_open_tickets(event_id, event_timezone='UTC'): tickets = Ticket.query.filter(Ticket.event_id == event_id).filter( Ticket.sales_start <= datetime.datetime.now(pytz.timezone(event_timezone)).replace(tzinfo=None)).filter( Ticket.sales_end >= datetime.datetime.now(pytz.timezone(event_timezone)).replace(tzinfo=None)) …
In this case, first current time is found out using timezone method in the timezone which is stored as a separate data field in the database. Since comparison cannot be done between aware and naive date-time. Hence once current date-time is found out in the user’s timezone, it is made naive using replace method which makes the aware date-time into naive again. Hence can be compared with the naive date-time stored already.