将Pandas DF列从UTC转换为美国东部时间,不带日期

Question

我有这个Pandas数据帧列:

                     time_UTC
0   2015-01-05 16:44:34+00:00
1   2015-08-11 16:44:38+00:00
2   2015-08-02 16:53:25+00:00
3   2015-08-17 16:53:25+00:00
4   2015-09-28 16:53:26+00:00
Name: time_UTC, dtype: datetime64[ns, UTC]

我使用以下方法将其从UTC转换为美国东部时区:

list_temp = []
    for row in df['time_UTC']:
        list_temp.append(Timestamp(row, tz = 'UTC').tz_convert('US/Eastern'))
    df['time_EST'] = list_temp

得到这个:

0   2015-01-05 11:44:34-05:00
1   2015-08-11 11:44:38-05:00
2   2015-08-02 11:53:25-05:00
3   2015-08-17 11:53:25-05:00
4   2015-09-28 11:53:26-05:00
Name: time_EST, dtype: datetime64[ns, US/Eastern]

现在,我需要删除条目的日期部分,以便我只得到时间.这就是我需要的:

0   11:44:34-05:00
1   11:44:38-05:00
2   11:53:25-05:00
3   11:53:25-05:00
4   11:53:26-05:00
Name: time_EST, dtype: datetime64[ns, US/Eastern]

尝试:

我试过这个:

print df['time_EST'].apply(lambda x: dt.time(x.hour,x.minute,x.second))

进行转换以便日期被删除,我只有时间.但它正在恢复到UTC时区.以下是上述命令的输出:

0    16:44:34
1    16:44:38
2    16:53:25
3    16:53:25
4    16:53:26
Name: time_EST, dtype: object

题:

有没有办法放弃日期并保持美国东部时间,而不会自动恢复到UTC？

编辑:

要重新创建问题,只需复制上面的第一个DataFrame并使用以下代码:

import pandas as pd
from pandas.lib import Timestamp
import datetime as dt
df = pd.read_clipboard()

然后从问题中复制剩余的代码行.对此问题的任何帮助将不胜感激.

Answer 1

您想用来strftime格式化字符串,还要注意矢量化日期操作:

df = pd.read_clipboard()
df.time_UTC = pd.to_datetime(df.time_UTC)
df['EST'] = (df.time_UTC.dt.tz_localize('UTC')
                        .tz_convert('US/Eastern')
                        .strftime("%H:%M:%S"))

In [41]: df
Out[41]:
                               time_UTC       EST
time_UTC
2016-02-15 16:44:34 2016-02-15 16:44:34  11:44:34
2016-02-15 16:44:38 2016-02-15 16:44:38  11:44:38
2016-02-15 16:53:25 2016-02-15 16:53:25  11:53:25
2016-02-15 16:53:25 2016-02-15 16:53:25  11:53:25
2016-02-15 16:53:26 2016-02-15 16:53:26  11:53:26