将 PDT/PST 时区列转换为 UTC 时区

Question

我有一个包含 PDT 和 PST 日期时间值的 pandas 列。示例：

PDT/PST
2021-10-29 00:18:38 PDT
2021-10-29 01:08:19 PDT
2021-11-08 19:43:58 PST
2021-11-08 19:56:01 PST

我需要将它们转换成 UTC 时区。示例：

UTC
2021-10-29 07:18:00

感谢简单的回答。

Answer 1

使用to_datetime with convert strings to datetimes by dateparser.parse:

import dateparser

df['PDT/PST'] = pd.to_datetime(df['PDT/PST'].apply(dateparser.parse), utc=True)
print (df)
                    PDT/PST
0 2021-10-29 07:18:38+00:00
1 2021-10-29 08:08:19+00:00
2 2021-11-09 03:43:58+00:00
3 2021-11-09 03:56:01+00:00

最后添加 Series.dt.tz_localize 与 None:

df['PDT/PST'] = (pd.to_datetime(df['PDT/PST'].apply(dateparser.parse), utc=True)
                   .dt.tz_localize(None))
print (df)
              PDT/PST
0 2021-10-29 07:18:38
1 2021-10-29 08:08:19
2 2021-11-09 03:43:58
3 2021-11-09 03:56:01

将 PST 和 PDT 替换为 -7 或 -8 的解决方案是：

df['PDT/PST'] = (pd.to_datetime(df['PDT/PST']
                                .replace({'PDT':'-07:00','PST':'-08:00'}, regex=True), utc=True)
                  .dt.tz_localize(None))
print (df)
              PDT/PST
0 2021-10-29 07:18:38
1 2021-10-29 08:08:19
2 2021-11-09 03:43:58
3 2021-11-09 03:56:01

Answer 2

另一种选择：dateutil 的解析器提供了 tzinfos；然后转换为 UTC。

import dateutil
pacific_tz = dateutil.tz.gettz("US/Pacific")

df['UTC'] = df['PDT/PST'].apply(dateutil.parser.parse,
                                tzinfos={'PST': pacific_tz,
                                         'PDT': pacific_tz}).dt.tz_convert('UTC')

df['UTC']

0   2021-10-29 07:18:38+00:00
1   2021-10-29 08:08:19+00:00
2   2021-11-09 03:43:58+00:00
3   2021-11-09 03:56:01+00:00
Name: UTC, dtype: datetime64[ns, UTC]

相关：Python strptime() and timezones?

现在您可以根据需要格式化为具有特定格式的字符串，例如

df['UTC'].dt.strftime('%Y-%m-%d %H:%M:%S')

0    2021-10-29 07:18:38
1    2021-10-29 08:08:19
2    2021-11-09 03:43:58
3    2021-11-09 03:56:01
Name: UTC, dtype: object

将 PDT/PST 时区列转换为 UTC 时区

Converting PDT/PST timezone column to UTC timezone

python

datetime

python-3.x

pandas

jupyter-notebook