在 pandas 数据框中高效地转换时区 [英] Efficiently convert timezones in pandas dataframe

问题描述

我有一个大熊猫数据框(数千万行)，其中包含一列UTC时间和时区.我想基于其他两个列创建一个包含行的本地时间的列.

I have a large pandas dataframe (tens of millions of rows) which includes a column for UTC time and the time zone. I want to create a column which contains the local time for the row, based on these two other columns.

我最初的尝试是使用 df.apply ，它在我正在测试的一个小样本上工作，但是非常慢，并且不足以处理整个数据:

My original attempt was using df.apply which worked on a small sample that I was testing on, but is very slow and isn't good enough to work on the whole data:

df['LoginTimeLocal'] = \ 
    df.apply(lambda row: row.LoginTimeUtc.tz_localize('UTC').tz_convert(row.TimeZoneCode))

这将导致添加新列，其中包含本地时间的日期时间以及时区信息.

This results in a new column being added which contains a datetime in local time, with timezone information.

我遇到了此答案，该答案提供了一种高效的矢量化方式来执行类似操作.我将这段代码重新用于执行我想要的操作，但它似乎仅在新列仅包含具有相同时区(或没有时区信息)的日期时才起作用.这是我的代码:

I came across this answer which provides an efficient, vectorized way to do something similar. I re-purposed this code to do what I want, but it appears to only work if the new column only contains dates with the same time zone (or no time zone information). Here's my code:

# localize all utc dates
df['LoginTimeUtc'] = df['LoginTimeUtc'].dt.tz_localize('UTC')

# initialize LoginTimeLocal column (probably not necessary)
df['LoginTimeLocal'] = df['LoginTimeUtc']

# for every time zone in the data
for tz in df.TimeZoneCode.unique():
    mask = (df.TimeZoneCode == tz)

    # make entries in a new column with converted timezone
    df.loc[mask, 'LoginTimeLocal'] = \ 
        df.loc[mask,'LoginTimeLocal'].dt.tz_convert(tz)

如果我对仅包含来自一个时区的日期的数据样本运行此操作(即 len(df.TimeZoneCode.unique())= 1 )，它会正常工作.一旦数据框中存在两个或多个时区，我就会得到一个 ValueError:不兼容或非tz意识的值.

If I run this on a sample of the data that only contains dates from one timezone (i.e., len(df.TimeZoneCode.unique()) = 1), it works fine. As soon as there are two or more timezones in the dataframe, I get a ValueError: incompatible or non tz-aware value.

有人可以看到这里出了什么问题吗?

Can anyone see what is going wrong here?

推荐答案

演示:

源DF:

In [11]: df
Out[11]:
             datetime         time_zone
0 2016-09-19 01:29:13    America/Bogota
1 2016-09-19 02:16:04  America/New_York
2 2016-09-19 01:57:54      Africa/Cairo
3 2016-09-19 11:00:00    America/Bogota
4 2016-09-19 12:00:00  America/New_York
5 2016-09-19 13:00:00      Africa/Cairo

解决方案:

In [12]: df['new'] =  df.groupby('time_zone')['datetime'] \
                        .transform(lambda x: x.dt.tz_localize(x.name))

In [13]: df
Out[13]:
             datetime         time_zone                 new
0 2016-09-19 01:29:13    America/Bogota 2016-09-19 06:29:13
1 2016-09-19 02:16:04  America/New_York 2016-09-19 06:16:04
2 2016-09-19 01:57:54      Africa/Cairo 2016-09-18 23:57:54
3 2016-09-19 11:00:00    America/Bogota 2016-09-19 16:00:00
4 2016-09-19 12:00:00  America/New_York 2016-09-19 16:00:00
5 2016-09-19 13:00:00      Africa/Cairo 2016-09-19 11:00:00

这篇关于在 pandas 数据框中高效地转换时区的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！