6

I am trying to dateutil.parser.parse() to parse the default str(datetime.datetime.now()) output using timezone-aware datetimes. However, parse() seems to lose the timezone information and replace it with the local time timezone. Below is the IPython output:

In [1]: from django.utils.timezone import now
In [3]: import dateutil
In [4]: t = now()
In [6]: print t
2014-07-14 08:51:49.123342+00:00

In [7]: st = unicode(t)
In [8]: print dateutil.parser.parse(st)
2014-07-14 08:51:49.123342+02:00

As far I understood dateutil does some heurestics when guessing the date format and it might go wrong here.

  • How to give exact datetime format for parsing timezone-aware datetimes?

  • Even better - if the format is known how to parse this datetime using only Python stdlib, without dateutil dependency?

mlissner
  • 17,359
  • 18
  • 106
  • 169
Mikko Ohtamaa
  • 82,057
  • 50
  • 264
  • 435
  • I am aware of `USE_TZ` - however this question is about parsing dates and `dateutil` behavior, not about Django. – Mikko Ohtamaa Jul 14 '14 at 09:17
  • 1
    Then why are you using `django.util.timezone`? – Burhan Khalid Jul 14 '14 at 09:19
  • I wanted to generate a datetime with timezone. That is kind of the easiest way. I think I was quite specific what I was asking - please tell me if there is anything in the question what can be improved. – Mikko Ohtamaa Jul 14 '14 at 09:21
  • 1
    The premise of your question is wrong - `datetime.datetime.now()` does not give you a timezone aware object; so when you do `str(datetime.datetime.now())`, there is no timezone information in the string. Is your question basically _"How to parse a string that represents a date and time with a timezone offset into a datetime object"_? – Burhan Khalid Jul 14 '14 at 09:24
  • Np. It happens to everyone :) – Mikko Ohtamaa Jul 14 '14 at 09:24
  • Actually I think it might be a bug regarding +00:00 handling. – Mikko Ohtamaa Jul 14 '14 at 09:26

1 Answers1

-3

timezone.now will give you an aware datetime object if USE_TZ is true. You don't need to parse it further.

In this example, USE_TZ is True, and the timezone is set to UTC:

>>> from django.utils import timezone as djtz
>>> i = djtz.now()
>>> type(i)
<type 'datetime.datetime'>
>>> i.tzinfo
<UTC>

As far as dateutil goes, it will parse your string correctly:

>>> from dateutil.parser import parse
>>> s = '2014-07-14 08:51:49.123342+00:00'
>>> parse(s).tzinfo
tzutc()
>>> z = parse(s)
>>> z
datetime.datetime(2014, 7, 14, 8, 51, 49, 123342, tzinfo=tzutc())

You can see its picking up the correct timezone (UTC in this case).

The default format specifiers only accept +0000 as the offset format with %z, or the three letter timezone name with %Z; but you cannot use this to parse, only to format:

>>> datetime.datetime.strptime('2014-07-14 08:51:49.123342+0000', '%Y-%m-%d %H:%M:%S.%f%z')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python2.7/_strptime.py", line 317, in _strptime
    (bad_directive, format))
ValueError: 'z' is a bad directive in format '%Y-%m-%d %H:%M:%S.%f%z'
>>> datetime.datetime.strftime(z, '%Z')
'UTC'
>>> datetime.datetime.strftime(z, '%z')
'+0000'
Burhan Khalid
  • 169,990
  • 18
  • 245
  • 284