python - 如何使用python中的GMT delta从时间戳中获取日期时间对象？

Question

我目前正在使用一个 android 应用程序收集时间戳，该应用程序为某些用户存储时区，例如“GMT+03:00”在线浏览我发现这不是一个正确的时区，因此在尝试构建日期时间时python中的对象使用

from datetime import datetime
from dateutil import tz
import pandas as pd
filename="data.csv"
data=pd.read_csv(filename)
[ datetime.fromtimestamp(data['timestamp'].iloc[i],
tz=tz.gettz(data['timezone'].iloc[i])) 
for i in range(data.shape[0]) ]

效果不好。例如，使用该日期时间对象作为索引来创建 Pandas 数据框以使用滚动窗口功能是行不通的。知道如何将“GMT+03:00”转换为适当的时区，或者以某种方式整合该信息以正确构建日期时间对象吗？

更新： 这是data['timestamps']的示例：

[1520719558.0, 1520719558.0, 1520719558.0, 1520719558.0, 1520719561.0, 1520719561.0, 1520719561.0, 1520719561.0, 1520719562.0, 1520719562.0]

还有data['timezone']的样本：

['GMT+03:00', 'GMT+03:00', 'GMT+03:00', 'GMT+03:00', 'GMT+03:00', 'GMT+03:00', 'GMT+03:00', 'GMT+03:00', 'GMT+03:00', 'GMT+03:00']

Answer 1

GMT 和 UTC 相同。您可以手动进行：编写一个函数来提取偏移量并返回datetime.timezone。

import datetime, re
def get_tz(s):
    '''Returns a datetime.timezone object.

    Uses regular expression to extract the UTC offset from s.
    Assumes s is in the form of "GMT+03:00" or "GMT-03:00".
    Does NOT have exception handling.

    '''
    pattern = r'GMT([+-])(\d{1,2}):(\d{1,2})'
    match = re.match(pattern, s)
    sign, hh, mm = match.groups()
    hh, mm = map(int, (hh, mm))
    t_delta = datetime.timedelta(hours=hh, minutes=mm)
    t_delta = t_delta * (1 if sign == '+' else -1)
    return datetime.timezone(t_delta)

用法：

>>> timestamp = 1520719558.0
>>> timezone = 'GMT+03:00'
>>> dt = datetime.datetime.fromtimestamp(timestamp, get_tz(timezone))
>>> dt.isoformat()
'2018-03-11T01:05:58+03:00'
>>> timezone = 'GMT-03:00'
>>> dt = datetime.datetime.fromtimestamp(timestamp, get_tz(timezone))
>>> dt.isoformat()
'2018-03-10T19:05:58-03:00'

Answer 2

#!/usr/bin/python3.5

import pandas as pd
import re
import datetime as dt

# From wwii solution
def get_tz(s):
    '''Returns a datetime.timezone object.

    Uses regular expression to extract the UTC offset from s.
    Assumes s is in the form of "GMT+03:00" or "GMT-03:00".
    Does NOT have exception handling.

    '''
    pattern = r'GMT([+-])(\d{1,2}):(\d{1,2})'
    match = re.match(pattern, s)
    sign, hh, mm = match.groups()
    hh, mm = map(int, (hh, mm))
    t_delta = dt.timedelta(hours=hh, minutes=mm)
    t_delta = t_delta * (1 if sign == '+' else -1)
    return dt.timezone(t_delta)

timestamps = [1520719558.0, 1520719558.0, 1520719558.0, 1520719558.0, 
              1520719561.0, 1520719561.0, 1520719561.0, 1520719561.0,
              1520719562.0, 1520719562.0]
timezones = ['GMT+03:00', 'GMT+03:00', 'GMT+03:00', 'GMT+03:00',
             'GMT+03:00', 'GMT+03:00', 'GMT+03:00', 'GMT+03:00',
             'GMT+03:00', 'GMT+03:00']
data = zip(timestamps, timezones)
data_df = pd.DataFrame(list(data), columns=['timestamp', 'timezone'])

# Converts timezone to date object
data_df['timezone'] = data_df['timezone'].apply(lambda x:get_tz(x))

# Adding a new-column in the dataframe with the datetime format requested
data_df['date_time'] = [dt.datetime.fromtimestamp(row['timestamp'], row['timezone'])  
                        for (_, row) in data_df[['timestamp', 'timezone']].iterrows()
                      ]