Convert timestamps with offset to datetime obj using strptime

I am trying to convert time-stamps of the format "2012-07-24T23:14:29-07:00" to datetime objects in python using strptime method. The problem is with the time offset at the end(-07:00). Without the offset i can successfully do

time_str = "2012-07-24T23:14:29"

time_obj=datetime.datetime.strptime(time_str,'%Y-%m-%dT%H:%M:%S')

But with the offset i tried

time_str = "2012-07-24T23:14:29-07:00"

time_obj=datetime.datetime.strptime(time_str,'%Y-%m-%dT%H:%M:%S-%z').

But it gives a Value error saying "z" is a bad directive.

Any ideas for a work around?


The Python 2 strptime() function indeed does not support the %z format for timezones (because the underlying time.strptime() function doesn't support it). You have two options:

  • Ignore the timezone when parsing with strptime:

    time_obj = datetime.datetime.strptime(time_str[:19], '%Y-%m-%dT%H:%M:%S')
    
  • use the dateutil module, it's parse function does deal with timezones:

    from dateutil.parser import parse
    time_obj = parse(time_str)
    

Quick demo on the command prompt:

>>> from dateutil.parser import parse
>>> parse("2012-07-24T23:14:29-07:00")
datetime.datetime(2012, 7, 24, 23, 14, 29, tzinfo=tzoffset(None, -25200))

You could also upgrade to Python 3.2 or newer, where timezone support has been improved to the point that %z would work, provided you remove the last : from the input, and the - from before the %z:

>>> import datetime
>>> time_str = "2012-07-24T23:14:29-07:00"
>>> datetime.datetime.strptime(time_str, '%Y-%m-%dT%H:%M:%S%z')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/mj/Development/Library/buildout.python/parts/opt/lib/python3.4/_strptime.py", line 500, in _strptime_datetime
    tt, fraction = _strptime(data_string, format)
  File "/Users/mj/Development/Library/buildout.python/parts/opt/lib/python3.4/_strptime.py", line 337, in _strptime
    (data_string, format))
ValueError: time data '2012-07-24T23:14:29-07:00' does not match format '%Y-%m-%dT%H:%M:%S%z'
>>> ''.join(time_str.rsplit(':', 1))
'2012-07-24T23:14:29-0700'
>>> datetime.datetime.strptime(''.join(time_str.rsplit(':', 1)), '%Y-%m-%dT%H:%M:%S%z')
datetime.datetime(2012, 7, 24, 23, 14, 29, tzinfo=datetime.timezone(datetime.timedelta(-1, 61200)))

In Python 3.7+:

from datetime import datetime

time_str = "2012-07-24T23:14:29-07:00"
dt_aware = datetime.fromisoformat(time_str)
print(dt_aware.isoformat('T'))
# -> 2012-07-24T23:14:29-07:00

In Python 3.2+:

from datetime import datetime

time_str = "2012-07-24T23:14:29-0700"
dt_aware = datetime.strptime(time_str, '%Y-%m-%dT%H:%M:%S%z')
print(dt_aware.isoformat('T'))
# -> 2012-07-24T23:14:29-07:00

Note: Before Python 3.7 this variant didn't support : in the -0700 part (both formats are allowed by rfc 3339). See datetime: add ability to parse RFC 3339 dates and times.

On older Python versions such as Python 2.7, you could parse the utc offset manually:

from datetime import datetime

time_str = "2012-07-24T23:14:29-0700"
# split the utc offset part
naive_time_str, offset_str = time_str[:-5], time_str[-5:]
# parse the naive date/time part
naive_dt = datetime.strptime(naive_time_str, '%Y-%m-%dT%H:%M:%S')
# parse the utc offset
offset = int(offset_str[-4:-2])*60 + int(offset_str[-2:])
if offset_str[0] == "-":
   offset = -offset
dt = naive_dt.replace(tzinfo=FixedOffset(offset))
print(dt.isoformat('T'))

where FixedOffset class is defined here.


ValueError: 'z' is a bad directive in format...

(note: I have to stick to python 2.7 in my case)

I have had a similar problem parsing commit dates from the output of git log --date=iso8601 which actually isn't the ISO8601 format (hence the addition of --date=iso8601-strict in a later version).

Since I am using django I can leverage the utilities there.

https://github.com/django/django/blob/master/django/utils/dateparse.py

>>> from django.utils.dateparse import parse_datetime
>>> parse_datetime('2013-07-23T15:10:59.342107+01:00')
datetime.datetime(2013, 7, 23, 15, 10, 59, 342107, tzinfo=+0100)

Instead of strptime you could use your own regular expression.


With python 3.5.2
To convert 26 Sep 2000 05:11:00 -0700

from datetime import datetime    
dt_obj = datetime.strptime("26 Sep 2000 05:11:00 -0700", '%d %b %Y %H:%M:%S %z')

To convert 2012-07-24T23:14:29 -0700

dt_obj = datetime.strptime('2012-07-24T23:14:29 -0700', '%Y-%m-%dT%H:%M:%S %z')

Python 3.5.2 doesn't support -07:00 time offset ':' should be removed