Python 实用宝典

Question 1

I have a python datetime instance that was created using datetime.utcnow() and persisted in database.

For display, I would like to convert the datetime instance retrieved from the database to local datetime using the default local timezone (i.e., as if the datetime was created using datetime.now()).

How can I convert the UTC datetime to a local datetime using only python standard library (e.g., no pytz dependency)?

It seems one solution would be to use datetime.astimezone( tz ), but how would you get the default local timezone?

Question 2

In Python 3.3+:

from datetime import datetime, timezone

def utc_to_local(utc_dt):
    return utc_dt.replace(tzinfo=timezone.utc).astimezone(tz=None)

In Python 2/3:

import calendar
from datetime import datetime, timedelta

def utc_to_local(utc_dt):
    # get integer timestamp to avoid precision lost
    timestamp = calendar.timegm(utc_dt.timetuple())
    local_dt = datetime.fromtimestamp(timestamp)
    assert utc_dt.resolution >= timedelta(microseconds=1)
    return local_dt.replace(microsecond=utc_dt.microsecond)

Using pytz (both Python 2/3):

import pytz

local_tz = pytz.timezone('Europe/Moscow') # use your local timezone name here
# NOTE: pytz.reference.LocalTimezone() would produce wrong result here

## You could use `tzlocal` module to get local timezone on Unix and Win32
# from tzlocal import get_localzone # $ pip install tzlocal

# # get local timezone    
# local_tz = get_localzone()

def utc_to_local(utc_dt):
    local_dt = utc_dt.replace(tzinfo=pytz.utc).astimezone(local_tz)
    return local_tz.normalize(local_dt) # .normalize might be unnecessary

Example

def aslocaltimestr(utc_dt):
    return utc_to_local(utc_dt).strftime('%Y-%m-%d %H:%M:%S.%f %Z%z')

print(aslocaltimestr(datetime(2010,  6, 6, 17, 29, 7, 730000)))
print(aslocaltimestr(datetime(2010, 12, 6, 17, 29, 7, 730000)))
print(aslocaltimestr(datetime.utcnow()))

Output

Python 3.3

2010-06-06 21:29:07.730000 MSD+0400
2010-12-06 20:29:07.730000 MSK+0300
2012-11-08 14:19:50.093745 MSK+0400

Python 2

2010-06-06 21:29:07.730000 
2010-12-06 20:29:07.730000 
2012-11-08 14:19:50.093911

pytz

2010-06-06 21:29:07.730000 MSD+0400
2010-12-06 20:29:07.730000 MSK+0300
2012-11-08 14:19:50.146917 MSK+0400

Note: it takes into account DST and the recent change of utc offset for MSK timezone.

I don’t know whether non-pytz solutions work on Windows.

Question 3

You can’t do it with only the standard library as the standard library doesn’t have any timezones. You need pytz or dateutil.

>>> from datetime import datetime
>>> now = datetime.utcnow()
>>> from dateutil import tz
>>> HERE = tz.tzlocal()
>>> UTC = tz.gettz('UTC')

The Conversion:
>>> gmt = now.replace(tzinfo=UTC)
>>> gmt.astimezone(HERE)
datetime.datetime(2010, 12, 30, 15, 51, 22, 114668, tzinfo=tzlocal())

Or well, you can do it without pytz or dateutil by implementing your own timezones. But that would be silly.

Question 4

You can’t do it with standard library. Using pytz module you can convert any naive/aware datetime object to any other time zone. Lets see some examples using Python 3.

Naive objects created through class method utcnow()

To convert a naive object to any other time zone, first you have to convert it into aware datetime object. You can use the replace method for converting a naive datetime object to an aware datetime object. Then to convert an aware datetime object to any other timezone you can use astimezone method.

The variable pytz.all_timezones gives you the list of all available time zones in pytz module.

import datetime,pytz

dtobj1=datetime.datetime.utcnow()   #utcnow class method
print(dtobj1)

dtobj3=dtobj1.replace(tzinfo=pytz.UTC) #replace method

dtobj_hongkong=dtobj3.astimezone(pytz.timezone("Asia/Hong_Kong")) #astimezone method
print(dtobj_hongkong)

Naive objects created through class method now()

Because now method returns current date and time, so you have to make the datetime object timezone aware first. The localize function converts a naive datetime object into a timezone-aware datetime object. Then you can use the astimezone method to convert it into another timezone.

dtobj2=datetime.datetime.now()

mytimezone=pytz.timezone("Europe/Vienna") #my current timezone
dtobj4=mytimezone.localize(dtobj2)        #localize function

dtobj_hongkong=dtobj4.astimezone(pytz.timezone("Asia/Hong_Kong")) #astimezone method
print(dtobj_hongkong)

Question 5

I think I figured it out: computes number of seconds since epoch, then converts to a local timzeone using time.localtime, and then converts the time struct back into a datetime…

EPOCH_DATETIME = datetime.datetime(1970,1,1)
SECONDS_PER_DAY = 24*60*60

def utc_to_local_datetime( utc_datetime ):
    delta = utc_datetime - EPOCH_DATETIME
    utc_epoch = SECONDS_PER_DAY * delta.days + delta.seconds
    time_struct = time.localtime( utc_epoch )
    dt_args = time_struct[:6] + (delta.microseconds,)
    return datetime.datetime( *dt_args )

It applies the summer/winter DST correctly:

>>> utc_to_local_datetime( datetime.datetime(2010, 6, 6, 17, 29, 7, 730000) )
datetime.datetime(2010, 6, 6, 19, 29, 7, 730000)
>>> utc_to_local_datetime( datetime.datetime(2010, 12, 6, 17, 29, 7, 730000) )
datetime.datetime(2010, 12, 6, 18, 29, 7, 730000)

Question 6

Building on Alexei’s comment. This should work for DST too.

import time
import datetime

def utc_to_local(dt):
    if time.localtime().tm_isdst:
        return dt - datetime.timedelta(seconds = time.altzone)
    else:
        return dt - datetime.timedelta(seconds = time.timezone)

Question 7

The standard Python library does not come with any tzinfo implementations at all. I’ve always considered this a surprising shortcoming of the datetime module.

The documentation for the tzinfo class does come with some useful examples. Look for the large code block at the end of the section.

Question 8

Python 3.9 adds the zoneinfo module so now it can be done as follows (stdlib only):

from zoneinfo import ZoneInfo
from datetime import datetime

utc_unaware = datetime(2020, 10, 31, 12)  # loaded from database
utc_aware = utc_unaware.replace(tzinfo=ZoneInfo('UTC'))  # make aware
local_aware = utc_aware.astimezone(ZoneInfo('localtime'))  # convert

Central Europe is 1 or 2 hours ahead of UTC, so local_aware is:

datetime.datetime(2020, 10, 31, 13, 0, tzinfo=backports.zoneinfo.ZoneInfo(key='localtime'))

as str:

2020-10-31 13:00:00+01:00

Windows has no system time zone database, so here an extra package is needed:

pip install tzdata

There is a backport to allow use in Python 3.6 to 3.8:

sudo pip install backports.zoneinfo

Then:

from backports.zoneinfo import ZoneInfo

Question 9

A simple (but maybe flawed) way that works in Python 2 and 3:

import time
import datetime

def utc_to_local(dt):
    return dt - datetime.timedelta(seconds = time.timezone)

Its advantage is that it’s trivial to write an inverse function

Question 10

The easiest way I have found is to get the time offset of where you are, then subtract that from the hour.

def format_time(ts,offset):
    if not ts.hour >= offset:
        ts = ts.replace(day=ts.day-1)
        ts = ts.replace(hour=ts.hour-offset)
    else:
        ts = ts.replace(hour=ts.hour-offset)
    return ts

This works for me, in Python 3.5.2.

Question 11

Here is another way to change timezone in datetime format (I know I wasted my energy on this but I didn’t see this page so I don’t know how) without min. and sec. cause I don’t need it for my project:

def change_time_zone(year, month, day, hour):
      hour = hour + 7 #<-- difference
      if hour >= 24:
        difference = hour - 24
        hour = difference
        day += 1
        long_months = [1, 3, 5, 7, 8, 10, 12]
        short_months = [4, 6, 9, 11]
        if month in short_months:
          if day >= 30:
            day = 1
            month += 1
            if month > 12:
              year += 1
        elif month in long_months:
          if day >= 31:
            day = 1
            month += 1
            if month > 12:
              year += 1
        elif month == 2:
          if not year%4==0:
            if day >= 29:
              day = 1
              month += 1
              if month > 12:
                year += 1
          else:
            if day >= 28:
              day = 1
              month += 1
              if month > 12:
                year += 1
      return datetime(int(year), int(month), int(day), int(hour), 00)

Question 12

This is a terrible way to do it but it avoids creating a definition. It fulfills the requirement to stick with the basic Python3 library.

# Adjust from UST to Eastern Standard Time (dynamic)
# df.my_localtime should already be in datetime format, so just in case
df['my_localtime'] = pd.to_datetime.df['my_localtime']

df['my_localtime'] = df['my_localtime'].dt.tz_localize('UTC').dt.tz_convert('America/New_York').astype(str)
df['my_localtime'] = pd.to_datetime(df.my_localtime.str[:-6])

Question 13

Use timedelta to switch between timezones. All you need is the offset in hours between timezones. Don’t have to fiddle with boundaries for all 6 elements of a datetime object. timedelta handles leap years, leap centuries, etc., too, with ease. You must first

from datetime import datetime, timedelta

Then if offset is the timezone delta in hours:

timeout = timein + timedelta(hours = offset)

where timein and timeout are datetime objects. e.g.

timein + timedelta(hours = -8)

converts from GMT to PST.

So, how to determine offset? Here is a simple function provided you only have a few possibilities for conversion without using datetime objects that are timezone “aware” which some other answers nicely do. A bit manual, but sometimes clarity is best.

def change_timezone(timein, timezone, timezone_out):
    '''
    changes timezone between predefined timezone offsets to GMT
    timein - datetime object
    timezone - 'PST', 'PDT', 'GMT' (can add more as needed)
    timezone_out - 'PST', 'PDT', 'GMT' (can add more as needed)
    ''' 
    # simple table lookup        
    tz_offset =  {'PST': {'GMT': 8, 'PDT': 1, 'PST': 0}, \
                  'GMT': {'PST': -8, 'PDT': -7, 'GMT': 0}, \
                  'PDT': {'GMT': 7, 'PST': -1, 'PDT': 0}}
    try:
        offset = tz_offset[timezone][timezone_out]
    except:
        msg = 'Input timezone=' + timezone + ' OR output time zone=' + \
            timezone_out + ' not recognized'
        raise DateTimeError(msg)

    return timein + timedelta(hours = offset)

After looking at the numerous answers and playing around with the tightest code I can think of (for now) it seems best that all applications, where time is important and mixed timezones must be accounted for, should make a real effort to make all datetime objects “aware”. Then it would seem the simplest answer is:

timeout = timein.astimezone(pytz.timezone("GMT"))

to convert to GMT for example. Of course, to convert to/from any other timezone you wish, local or otherwise, just use the appropriate timezone string that pytz understands (from pytz.all_timezones). Daylight savings time is then also taken into account.

Question 14

How can I convert a DataFrame column of strings (in dd/mm/yyyy format) to datetimes?

Question 15

The easiest way is to use to_datetime:

df['col'] = pd.to_datetime(df['col'])

It also offers a dayfirst argument for European times (but beware this isn’t strict).

Here it is in action:

In [11]: pd.to_datetime(pd.Series(['05/23/2005']))
Out[11]:
0   2005-05-23 00:00:00
dtype: datetime64[ns]

You can pass a specific format:

In [12]: pd.to_datetime(pd.Series(['05/23/2005']), format="%m/%d/%Y")
Out[12]:
0   2005-05-23
dtype: datetime64[ns]

Question 16

If your date column is a string of the format ‘2017-01-01’ you can use pandas astype to convert it to datetime.

df['date'] = df['date'].astype('datetime64[ns]')

or use datetime64[D] if you want Day precision and not nanoseconds

print(type(df_launath['date'].iloc[0]))

yields

<class 'pandas._libs.tslib.Timestamp'> the same as when you use pandas.to_datetime

You can try it with other formats then ‘%Y-%m-%d’ but at least this works.

Question 17

You can use the following if you want to specify tricky formats:

df['date_col'] =  pd.to_datetime(df['date_col'], format='%d/%m/%Y')

More details on format here:

Question 18

If you have a mixture of formats in your date, don’t forget to set infer_datetime_format=True to make life easier

df['date'] = pd.to_datetime(df['date'], infer_datetime_format=True)

Source: pd.to_datetime

or if you want a customized approach:

def autoconvert_datetime(value):
    formats = ['%m/%d/%Y', '%m-%d-%y']  # formats to try
    result_format = '%d-%m-%Y'  # output format
    for dt_format in formats:
        try:
            dt_obj = datetime.strptime(value, dt_format)
            return dt_obj.strftime(result_format)
        except Exception as e:  # throws exception when format doesn't match
            pass
    return value  # let it be if it doesn't match

df['date'] = df['date'].apply(autoconvert_datetime)

问题：仅使用python标准库将python UTC日期时间转换为本地日期时间？

回答 0

例

输出量

Example

Output

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

回答 8

回答 9

回答 10

回答 11

问题：将DataFrame列类型从字符串转换为日期时间，格式为dd / mm / yyyy

回答 0

回答 1

回答 2

回答 3

问题：在Python中以YYYY-MM-DD获取今天的日期？

回答 0

回答 1

回答 2

回答 3

回答 4

潘达斯

摆

PANDAS

PENDULUM

回答 5

回答 6

回答 7

有趣好用的Python教程