我需要将RFC 3339字符串(如“2008-09-03T20:56:55.450686Z”)解析为Python的datetime类型。

我在Python标准库中找到了strptime,但它不是很方便。

最好的方法是什么?


当前回答

感谢Mark Amery的出色回答,我设计了一个函数来解释所有可能的ISO日期时间格式:

class FixedOffset(tzinfo):
    """Fixed offset in minutes: `time = utc_time + utc_offset`."""
    def __init__(self, offset):
        self.__offset = timedelta(minutes=offset)
        hours, minutes = divmod(offset, 60)
        #NOTE: the last part is to remind about deprecated POSIX GMT+h timezones
        #  that have the opposite sign in the name;
        #  the corresponding numeric value is not used e.g., no minutes
        self.__name = '<%+03d%02d>%+d' % (hours, minutes, -hours)
    def utcoffset(self, dt=None):
        return self.__offset
    def tzname(self, dt=None):
        return self.__name
    def dst(self, dt=None):
        return timedelta(0)
    def __repr__(self):
        return 'FixedOffset(%d)' % (self.utcoffset().total_seconds() / 60)
    def __getinitargs__(self):
        return (self.__offset.total_seconds()/60,)

def parse_isoformat_datetime(isodatetime):
    try:
        return datetime.strptime(isodatetime, '%Y-%m-%dT%H:%M:%S.%f')
    except ValueError:
        pass
    try:
        return datetime.strptime(isodatetime, '%Y-%m-%dT%H:%M:%S')
    except ValueError:
        pass
    pat = r'(.*?[+-]\d{2}):(\d{2})'
    temp = re.sub(pat, r'\1\2', isodatetime)
    naive_date_str = temp[:-5]
    offset_str = temp[-5:]
    naive_dt = datetime.strptime(naive_date_str, '%Y-%m-%dT%H:%M:%S.%f')
    offset = int(offset_str[-4:-2])*60 + int(offset_str[-2:])
    if offset_str[0] == "-":
        offset = -offset
    return naive_dt.replace(tzinfo=FixedOffset(offset))

其他回答

如果使用熊猫,我可以向熊猫推荐时间戳。在那里你可以

ts_1 = pd.Timestamp('2020-02-18T04:27:58.000Z')    
ts_2 = pd.Timestamp('2020-02-18T04:27:58.000')

Rant:令人难以置信的是,我们仍然需要担心2021的日期字符串解析等问题。

如今,Arrow还可以作为第三方解决方案:

>>> import arrow
>>> date = arrow.get("2008-09-03T20:56:35.450686Z")
>>> date.datetime
datetime.datetime(2008, 9, 3, 20, 56, 35, 450686, tzinfo=tzutc())

最初我尝试了:

from operator import neg, pos
from time import strptime, mktime
from datetime import datetime, tzinfo, timedelta

class MyUTCOffsetTimezone(tzinfo):
    @staticmethod
    def with_offset(offset_no_signal, signal):  # type: (str, str) -> MyUTCOffsetTimezone
        return MyUTCOffsetTimezone((pos if signal == '+' else neg)(
            (datetime.strptime(offset_no_signal, '%H:%M') - datetime(1900, 1, 1))
          .total_seconds()))

    def __init__(self, offset, name=None):
        self.offset = timedelta(seconds=offset)
        self.name = name or self.__class__.__name__

    def utcoffset(self, dt):
        return self.offset

    def tzname(self, dt):
        return self.name

    def dst(self, dt):
        return timedelta(0)


def to_datetime_tz(dt):  # type: (str) -> datetime
    fmt = '%Y-%m-%dT%H:%M:%S.%f'
    if dt[-6] in frozenset(('+', '-')):
        dt, sign, offset = strptime(dt[:-6], fmt), dt[-6], dt[-5:]
        return datetime.fromtimestamp(mktime(dt),
                                      tz=MyUTCOffsetTimezone.with_offset(offset, sign))
    elif dt[-1] == 'Z':
        return datetime.strptime(dt, fmt + 'Z')
    return datetime.strptime(dt, fmt)

但这在负时区不起作用。然而,我在Python 3.7.3中工作得很好:

from datetime import datetime


def to_datetime_tz(dt):  # type: (str) -> datetime
    fmt = '%Y-%m-%dT%H:%M:%S.%f'
    if dt[-6] in frozenset(('+', '-')):
        return datetime.strptime(dt, fmt + '%z')
    elif dt[-1] == 'Z':
        return datetime.strptime(dt, fmt + 'Z')
    return datetime.strptime(dt, fmt)

一些测试注意到,输出只在微秒的精度上有所不同。我的机器精度达到6位数,但YMMV:

for dt_in, dt_out in (
        ('2019-03-11T08:00:00.000Z', '2019-03-11T08:00:00'),
        ('2019-03-11T08:00:00.000+11:00', '2019-03-11T08:00:00+11:00'),
        ('2019-03-11T08:00:00.000-11:00', '2019-03-11T08:00:00-11:00')
    ):
    isoformat = to_datetime_tz(dt_in).isoformat()
    assert isoformat == dt_out, '{} != {}'.format(isoformat, dt_out)

Django的parse_datetime()函数支持UTC偏移的日期:

parse_datetime('2016-08-09T15:12:03.65478Z') =
datetime.datetime(2016, 8, 9, 15, 12, 3, 654780, tzinfo=<UTC>)

因此,它可以用于解析整个项目中字段中的ISO 8601日期:

from django.utils import formats
from django.forms.fields import DateTimeField
from django.utils.dateparse import parse_datetime

class DateTimeFieldFixed(DateTimeField):
    def strptime(self, value, format):
        if format == 'iso-8601':
            return parse_datetime(value)
        return super().strptime(value, format)

DateTimeField.strptime = DateTimeFieldFixed.strptime
formats.ISO_INPUT_FORMATS['DATETIME_INPUT_FORMATS'].insert(0, 'iso-8601')

Python>=3.11

fromsoformat现在直接解析Z:

from datetime import datetime

s = "2008-09-03T20:56:35.450686Z"

datetime.fromisoformat(s)
datetime.datetime(2008, 9, 3, 20, 56, 35, 450686, tzinfo=datetime.timezone.utc)

Python 3.7到3.10

一个注释中的简单选项:将“Z”替换为“+00:00”-并使用fromsoformat:

from datetime import datetime

s = "2008-09-03T20:56:35.450686Z"

datetime.fromisoformat(s.replace('Z', '+00:00'))
# datetime.datetime(2008, 9, 3, 20, 56, 35, 450686, tzinfo=datetime.timezone.utc)

为什么更喜欢来自同一格式?

虽然strptime的%z可以将“z”字符解析为UTC,但fromsoformat的速度要快~x40(另请参阅:更快的strptime):

%timeit datetime.fromisoformat(s.replace('Z', '+00:00'))
388 ns ± 48.3 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)

%timeit dateutil.parser.isoparse(s)
11 µs ± 1.05 µs per loop (mean ± std. dev. of 7 runs, 100000 loops each)

%timeit datetime.strptime(s, '%Y-%m-%dT%H:%M:%S.%f%z')
15.8 µs ± 1.32 µs per loop (mean ± std. dev. of 7 runs, 100000 loops each)

%timeit dateutil.parser.parse(s)
87.8 µs ± 8.54 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)

(Windows 10上的Python 3.9.12 x64)