I have a few records inside a column which represent either EST or EDT Time. I need to convert these times to GMT time. The format of the time are:
10/1/2010 0:0:0
10/1/2010 0:6:0
...
10/1/2010 23:54:0
...
10/3/2010 0:0:0
...
Can someone help me out here? thanks
The easiest, most reliable way I know to convert between timezones is to use the third-party pytz module:
import pytz
import datetime as dt
utc=pytz.utc
eastern=pytz.timezone('US/Eastern')
fmt='%Y-%m-%d %H:%M:%S %Z%z'
text='''\
10/1/2010 0:0:0
10/1/2010 0:6:0
10/1/2010 23:54:0
10/3/2010 0:0:0
'''
for datestring in text.splitlines():
date=dt.datetime.strptime(datestring,"%m/%d/%Y %H:%M:%S")
date_eastern=eastern.localize(date,is_dst=None)
date_utc=date_eastern.astimezone(utc)
print(date_utc.strftime(fmt))
yields:
2010-10-01 04:00:00 UTC+0000
2010-10-01 04:06:00 UTC+0000
2010-10-02 03:54:00 UTC+0000
2010-10-03 04:00:00 UTC+0000
Note however, your data does not specify if the datetime is in the EST or EDT timezone. There are some times which are ambiguous when you don't specify EST or EDT. For example, '10/27/2002 1:30:00' would be ambiguous:
>>> eastern.localize(datetime(2002, 10, 27, 1, 30, 00), is_dst=None)
AmbiguousTimeError: 2002-10-27 01:30:00
since this time happened twice due to Daylight Savings Time. Also some datetimes, like 2002-04-07 02:30:00, are nonexistent. See this link for a discussion of these and even more bizarre issues that arise when dealing with localtimes.
If you are willing to overlook these knotty corner cases, and if your machine is setup in the local timezone (e.g. EST/EDT),
there is a way to convert between the local and UTC timezones which does
not require the installation of pytz
. The idea is to convert the datetime --> timetuple --> timestamp --> UTC datetime. The chain of conversions is done with
dt.datetime.utcfromtimestamp(time.mktime(date.timetuple()))
For example:
import time
import datetime as dt
import pytz
utc=pytz.utc
eastern=pytz.timezone('US/Eastern')
fmt='%Y-%m-%d %H:%M:%S %Z%z'
text='''\
10/1/2010 0:0:0
10/1/2010 0:6:0
10/1/2010 23:54:0
10/3/2010 0:0:0
3/13/2011 1:55:0
3/13/2011 3:00:0
'''
for datestring in text.splitlines():
date=dt.datetime.strptime(datestring,"%m/%d/%Y %H:%M:%S")
date_est=eastern.localize(date,is_dst=None)
date_utc=date_est.astimezone(utc)
date_utc2=dt.datetime.utcfromtimestamp(time.mktime(date.timetuple()))
print('{d} --> {d_utc} {d_utc2}'.format(
d=date.strftime(fmt),
d_utc=date_utc.strftime(fmt),
d_utc2=date_utc2.strftime(fmt),
))
assert date_utc.hour == date_utc2.hour
yields
2010-10-01 00:00:00 EDT-0400 --> 2010-10-01 04:00:00 UTC+0000 2010-10-01 04:00:00
2010-10-01 00:06:00 EDT-0400 --> 2010-10-01 04:06:00 UTC+0000 2010-10-01 04:06:00
2010-10-01 23:54:00 EDT-0400 --> 2010-10-02 03:54:00 UTC+0000 2010-10-02 03:54:00
2010-10-03 00:00:00 EDT-0400 --> 2010-10-03 04:00:00 UTC+0000 2010-10-03 04:00:00
2011-03-13 01:55:00 EST-0500 --> 2011-03-13 06:55:00 UTC+0000 2011-03-13 06:55:00
2011-03-13 03:00:00 EDT-0400 --> 2011-03-13 07:00:00 UTC+0000 2011-03-13 07:00:00
The last two dates tested above show the conversion works correctly even with times close to the switch between EST and EDT.
In summary, using the alternate method (without pytz), here is how to convert datetime objects representing local time to datetime objects representing GMT time, and vice versa:
In [83]: import datetime as dt
In [84]: import time
In [85]: import calendar
In [86]: date=dt.datetime(2010,12,1,0,0,0)
In [87]: date
Out[87]: datetime.datetime(2010, 12, 1, 0, 0)
In [88]: date_utc=dt.datetime.utcfromtimestamp(time.mktime(date.timetuple()))
In [89]: date_utc
Out[89]: datetime.datetime(2010, 12, 1, 5, 0)
In [90]: date_local=dt.datetime.fromtimestamp(calendar.timegm(date_utc.timetuple()))
In [91]: date_local
Out[91]: datetime.datetime(2010, 12, 1, 0, 0)