python Pandas : Join on unique column values and concatenate

Question

我有三个 Pandas 数据帧，df1、df2、 和 df3，如下所示:

import pandas as pd
import numpy as np
df1 = pd.DataFrame({'id' : ['one', 'two', 'three'], 'score': [56, 45, 78]})
df2 = pd.DataFrame({'id' : ['one', 'five', 'four'], 'score': [35, 81, 90]})
df3 = pd.DataFrame({'id' : ['five', 'two', 'six'], 'score': [23, 66, 42]})

我如何根据 id 连接这些数据框，然后将它们的列连接在一起？所需的输出如下:

#join_and_concatenate by id:

id   score(df1)  score(df2)  score(df3)
one    56            35         NaN
two    45            NaN        66
three  78            NaN        NaN
four   NaN           90         NaN
five   NaN           81         23
six    NaN           NaN        42

我找到了一个相关的 page谈到 merge()、concatenate() 和 join() 但我不确定其中任何一个都能满足我的要求。

Answer 1

concat 可能有更好的方法，但这应该可行:

In [48]: pd.merge(df1, df2, how='outer', on='id').merge(df3, how='outer', on='id')
Out[48]: 
      id  score_x  score_y  score
0    one       56       35    NaN
1    two       45      NaN     66
2  three       78      NaN    NaN
3   five      NaN       81     23
4   four      NaN       90    NaN
5    six      NaN      NaN     42

[6 rows x 4 columns]

得到你想要的答案:

In [54]: merged = pd.merge(df1, df2, how='outer', on='id').merge(df3, how='outer', on='id')

In [55]: merged.set_index('id').rename(columns={'score_x': 'score(df1)', 'score_y': 'score(df2)
', 'score': 'score(df3)'})
Out[55]: 
       score(df1)  score(df2)  score(df3)
id                                       
one            56          35         NaN
two            45         NaN          66
three          78         NaN         NaN
five          NaN          81          23
four          NaN          90         NaN
six           NaN         NaN          42

[6 rows x 3 columns]