Question

我知道这个问题已经被问过几次了，但是我似乎无法使其正常工作。我想将等于特定值的所有行值复制到单独的列。使用下面的df，当np.nan位于Member中时，我想跨np.nan复制到Group。

import pandas as pd
import numpy as np

df = pd.DataFrame({
    'Member' : ['C1','C2','C3','C4','C5',np.nan,np.nan],                 
    'Min' : [5,5,5,5,5,5,5],    
    'Max' : [10,10,10,10,10,10,10],                 
    'Group' : [26,26,26,26,26,26,26],                        
    })

# Attempt 1
df['Group'] = np.where(df['Member'] == np.nan, df['Group'], df['Member'])

# Attempt 2
df['Group']  = df.apply(lambda x: x['Member'] if x['Group'] == np.nan else x['Group'], axis=1)

这两种尝试均无效。预期的输出是：

  Member  Min  Max  Group
0  C1     5    10   26   
1  C2     5    10   26   
2  C3     5    10   26   
3  C4     5    10   26   
4  C5     5    10   26   
5  NaN    5    10   Nan
6  NaN    5    10   Nan

Answer 1

代码具有特殊功能的解决方案，用于Series.isna测试缺失值，并将Member与Group中的np.where列交换：

df['Group'] = np.where(df['Member'].isna(), df['Member'], df['Group'])

#alternative with notna
#df['Group'] = np.where(df['Member'].notna(), df['Group'], df['Member'])

#alternative with loc
#df.loc[df['Member'].isna(), 'Group'] = df['Member']
print (df)
  Member  Min  Max Group
0     C1    5   10    26
1     C2    5   10    26
2     C3    5   10    26
3     C4    5   10    26
4     C5    5   10    26
5    NaN    5   10   NaN
6    NaN    5   10   NaN

其中col等于值复制到单独的col-熊猫

1 个答案: