将多个数据框组合成多索引列的数据框

Question

假设我有 3 个数据框，

df_1 代表收入

Year        TSLA    MSFT     AVY
2019         851     200   112.8
2018         725     150    92.6

df_2 代表其他一些收入

Year        TSLA    MSFT     AVY
2019          10      13      17
2018          11      14      18

和df_3 代表费用

Year        TSLA    MSFT     AVY
2019         110     213     317
2018         111     214     418

获取以下df的代码可能是什么？

       TSLA                             MSFT                             AVY
Year   revenues other_revenues expenses revenues other_revenues expenses revenues other_revenues expenses
2019        851             10      110      200             13      213    112.8             17      317
2018        725             11      111      150             14      214     92.6             18      418

多索引形式的列在哪里？

谢谢

Answer 1

使用concat 与DataFrame.swaplevel 和DataFrame.sort_index 对MultiIndex 进行排序：

#if not Year is index first create it
L = [x.set_index('Year') for x in [df_1, df_2, df_3]]
df = (pd.concat(L, 
               axis=1, 
               keys=('evenues', 'other_revenues', 'expenses'))
        .swaplevel(1, 0, axis=1)
        .sort_index(axis=1))
print (df)
         AVY                            MSFT                            TSLA  \
     evenues expenses other_revenues evenues expenses other_revenues evenues   
Year                                                                           
2019   112.8      317             17     200      213             13     851   
2018    92.6      418             18     150      214             14     725   

                              
     expenses other_revenues  
Year                          
2019      110             10  
2018      111             11  

编辑：对于原始订单中的订单，请从MultiIndex 的第一级唯一值中添加MultiIndex.from_product DataFrame.reindex：

sub = ['revenues', 'other_revenues', 'expenses']
L = [x.set_index('Year') for x in [df_1, df_2, df_3]]
df = (pd.concat(L, 
               axis=1, 
               keys=sub)
        .swaplevel(1, 0, axis=1))

mux = pd.MultiIndex.from_product([df.columns.levels[0], sub])
df = df.reindex(mux, axis=1)
        
print (df)
         TSLA                             MSFT                          \
     revenues other_revenues expenses revenues other_revenues expenses   
Year                                                                     
2019      851             10      110      200             13      213   
2018      725             11      111      150             14      214   

          AVY                          
     revenues other_revenues expenses  
Year                                   
2019    112.8             17      317  
2018     92.6             18      418