python - 如何添加根据现有日期列和标准列计算的数据框日期列？

Question

我有以下数据框。
日期是 dtype='datetime64[ns, UTC]'

    announce_date  announce_time
0   2013-01-23     After Market Close
1   2013-04-23     Before Market Open
2   2013-07-20     After Market Close

如何创建一个新的偏移列
如果 announce_time 是 “After Market Close”，则 announce_date 减 1。

    announce_date  announce_time          impact_date
0   2013-01-23     After Market Close     2013-01-24
1   2013-04-23     Before Market Open     2013-04-23
2   2013-07-20     After Market Close     2013-07-21

我可以做一个偏移日期列，但还不知道如何应用条件。

import pandas as pd

df['impact_date'] = df['announce_date'] + pd.DateOffset(days=1)

Answer 1

生成一个 bool 掩码并使用 pd.to_timedelta 从中生成偏移量。

# Convert "announce_date" to datetime if not already done.
# df['announce_date'] = pd.to_datetime(df['announce_date'], errors='coerce')
# Now calculate the offset.
offset = pd.to_timedelta(
    df['announce_time'].eq("After Market Close").astype(int), unit='D')    
df['impact_date'] = df['announce_date'] + offset

print(df)
  announce_date       announce_time impact_date
0    2013-01-23  After Market Close  2013-01-24
1    2013-04-23  Before Market Open  2013-04-23
2    2013-07-20  After Market Close  2013-07-21