Parse timezone abbreviation to UTC [duplicate]

问题:

This question already has answers here:

Parsing date/time string with timezone abbreviated name in Python? (5 answers)

Closed 5 years ago.

How can I convert a date time string of the form Feb 25 2010, 16:19:20 CET to the unix epoch?

Currently my best approach is to use time.strptime() is this:

def to_unixepoch(s):
    # ignore the time zone in strptime
    a = s.split()
    b = time.strptime(" ".join(a[:-1]) + " UTC", "%b %d %Y, %H:%M:%S %Z")
    # this puts the time_tuple(UTC+TZ) to unixepoch(UTC+TZ+LOCALTIME)
    c = int(time.mktime(b))
    # UTC+TZ
    c -= time.timezone
    # UTC
    c -= {"CET": 3600, "CEST": 2 * 3600}[a[-1]]
    return c

I see from other questions that it might be possible to use calendar.timegm(), and pytz among others to simplify this, but these don't handle the abbreviated time zones.

I'd like a solution that requires minimal excess libraries, I like to keep to the standard library as much as possible.

回答1:

The Python standard library does not really implement time zones. You should use python-dateutil. It provides useful extensions to the standard datetime module including a time zones implementation and a parser.

You can convert time zone aware datetime objects to UTC with .astimezone(dateutil.tz.tzutc()). For the current time as a timezone aware datetime object, you can use datetime.datetime.utcnow().replace(tzinfo=dateutil.tz.tzutc()).

import dateutil.tz

cet = dateutil.tz.gettz('CET')

cesttime = datetime.datetime(2010, 4, 1, 12, 57, tzinfo=cet)
cesttime.isoformat()
'2010-04-01T12:57:00+02:00'

cettime = datetime.datetime(2010, 1, 1, 12, 57, tzinfo=cet)
cettime.isoformat() 
'2010-01-01T12:57:00+01:00'

# does not automatically parse the time zone portion
dateutil.parser.parse('Feb 25 2010, 16:19:20 CET')\
    .replace(tzinfo=dateutil.tz.gettz('CET'))

Unfortunately this technique will be wrong during the repeated daylight savings time hour.