Pandas中如何优雅按条件用两列最大值填充Plan Date列？

Got it, let's tackle this problem with a clean, pandas-native approach that's way more elegant than using np.where(). First, let's make sure we're working with a properly formatted DataFrame (I converted the nanosecond timestamps in Plan Date to proper datetime objects for consistency):

import pandas as pd
import numpy as np

# Your original data converted to a usable DataFrame
df = pd.DataFrame({
    "Plan Date": {23186: np.nan, 18533: np.nan, 17980: np.nan, 28559: np.nan, 
                  22587: 1572480000000000000, 22522: 1580428800000000000, 
                  16639: 1572480000000000000, 14304: 1572480000000000000},
    "Support Date": {23186: pd.Timestamp('2020-12-31'), 
                     18533: pd.Timestamp('2022-03-31'), 
                     17980: pd.Timestamp('2022-03-31'), 
                     28559: pd.Timestamp('2019-02-28'), 
                     22587: pd.Timestamp('2019-07-31'), 
                     22522: pd.Timestamp('2019-07-31'), 
                     16639: pd.Timestamp('2019-10-31'), 
                     14304: pd.Timestamp('2019-10-31')},
    "Plan Deadline": {23186: pd.NaT, 18533: pd.NaT, 17980: pd.NaT, 28559: pd.NaT, 
                      22587: pd.Timestamp('2019-10-31'), 
                      22522: pd.Timestamp('2020-01-31'), 
                      16639: pd.Timestamp('2019-10-31'), 
                      14304: pd.Timestamp('2019-10-31')}
})

# Convert nanosecond timestamps in Plan Date to datetime
df['Plan Date'] = pd.to_datetime(df['Plan Date'])

The Cleanest Approach: Use loc + max(axis=1) #

Instead of clunky np.where() logic, we can leverage pandas' native indexing and row-wise max calculation to achieve exactly what you need:

# Only fill Plan Date where Plan Deadline is not missing
df.loc[df['Plan Deadline'].notna(), 'Plan Date'] = df[['Support Date', 'Plan Deadline']].max(axis=1)

Why This Works: #

• Precise Indexing: df.loc[df['Plan Deadline'].notna(), 'Plan Date'] targets exactly the rows and column we need to modify—no unnecessary operations on the entire DataFrame.
• Row-Wise Max: df[['Support Date', 'Plan Deadline']].max(axis=1) calculates the maximum date for each row across the two columns, which is exactly what we want to put into Plan Date.
• Readability: The code reads like plain English, making it easy for anyone to understand the logic at a glance.

Result After Processing: #

Plan Date Support Date Plan Deadline
23186        NaT   2020-12-31           NaT
18533        NaT   2022-03-31           NaT
17980        NaT   2022-03-31           NaT
28559        NaT   2019-02-28           NaT
22587 2019-10-31   2019-07-31    2019-10-31
22522 2020-01-31   2019-07-31    2020-01-31
16639 2019-10-31   2019-10-31    2019-10-31
14304 2019-10-31   2019-10-31    2019-10-31

Bonus: Non-Destructive Version #

If you don't want to modify the original DataFrame, just create a copy first:

new_df = df.copy()
new_df.loc[new_df['Plan Deadline'].notna(), 'Plan Date'] = new_df[['Support Date', 'Plan Deadline']].max(axis=1)

内容的提问来源于stack exchange，提问作者Bartek Malysz