Python/Pandas - 我怎样才能只保留外连接的父数据框？

Question

实际数据有更多的字段，在 df1 上更少，在 df2 上更多。有些列的名称略有不同。

# intialise data of lists.
data1 = {'NameA':['Tom', 'Nick', 'Krish', 'Jack'],
        'AgeA':[20, 21, 19, 18]}
data2 = {'NameB':['Tom', 'Nick', 'C', 'D'],
        'AgeB':[20, 21, 3, 4]}

# Create DataFrame
df1 = pd.DataFrame(data1)
df2 = pd.DataFrame(data2)
list = [df1, df2]

df1 = pd.merge(df1,df2,how='left',left_on=['NameA','AgeA'],right_on=['NameB','AgeB'])
print(df1)

输出 =

   NameA  AgeA NameB  AgeB
0    Tom    20   Tom  20.0
1   Nick    21  Nick  21.0
2  Krish    19   NaN   NaN
3   Jack    18   NaN   NaN

预期 =

我在实现左连接时遇到了麻烦，只留下了 Pandas/Python 的父表。有没有人有一些指示？谢谢。

Answer 1

在merge 中使用left join 和参数indicator 的解决方案：

df3 = pd.merge(df1,df2,how='left',left_on=['NameA','AgeA'],right_on=['NameB','AgeB'], indicator=True)
print(df3)
   NameA  AgeA NameB  AgeB     _merge
0    Tom    20   Tom  20.0       both
1   Nick    21  Nick  21.0       both
2  Krish    19   NaN   NaN  left_only
3   Jack    18   NaN   NaN  left_only

df = df3.loc[df3['_merge'].eq('left_only'), df1.columns]
print (df)
   NameA  AgeA
2  Krish    19
3   Jack    18

外连接解决方​​案：

df3 = pd.merge(df1,df2,how='outer',left_on=['NameA','AgeA'],right_on=['NameB','AgeB'], indicator=True)
print(df3)
   NameA  AgeA NameB  AgeB      _merge
0    Tom  20.0   Tom  20.0        both
1   Nick  21.0  Nick  21.0        both
2  Krish  19.0   NaN   NaN   left_only
3   Jack  18.0   NaN   NaN   left_only
4    NaN   NaN     C   3.0  right_only
5    NaN   NaN     D   4.0  right_only

df = df3.loc[df3['_merge'].eq('left_only'), df1.columns]
print (df)
   NameA  AgeA
2  Krish  19.0
3   Jack  18.0