合并两个具有分层列的数据框

Question

这是我第一次在 Pandas 中使用多索引，我需要一些帮助来将两个数据框与分层列合并。 这是我的两个数据框：

col_index = pd.MultiIndex.from_product([['a', 'b', 'c'], ['w', 'x']])
df1 = pd.DataFrame(np.ones([4,6]),columns=col_index, index=range(4))

     a         b         c     
     w    x    w    x    w    x
0  1.0  1.0  1.0  1.0  1.0  1.0
1  1.0  1.0  1.0  1.0  1.0  1.0
2  1.0  1.0  1.0  1.0  1.0  1.0
3  1.0  1.0  1.0  1.0  1.0  1.0

df2 = pd.DataFrame(np.zeros([2,6]),columns=col_index, index=range(2))

     a         b         c     
     w    x    w    x    w    x
0  0.0  0.0  0.0  0.0  0.0  0.0
1  0.0  0.0  0.0  0.0  0.0  0.0

当我使用合并方法时，我得到以下结果：

pd.merge(df1,df2, how='left', suffixes=('', '_2'), left_index = True, right_index= True ))

     a         b         c       a_2       b_2       c_2     
     w    x    w    x    w    x    w    x    w    x    w    x
0  1.0  1.0  1.0  1.0  1.0  1.0  0.0  0.0  0.0  0.0  0.0  0.0
1  1.0  1.0  1.0  1.0  1.0  1.0  0.0  0.0  0.0  0.0  0.0  0.0
2  1.0  1.0  1.0  1.0  1.0  1.0  NaN  NaN  NaN  NaN  NaN  NaN
3  1.0  1.0  1.0  1.0  1.0  1.0  NaN  NaN  NaN  NaN  NaN  NaN

但我想在较低级别合并两个数据帧，后缀在 ['w', 'x'] 上生效，如下所示：

     a                   b                   c               
     w  w_2    x  x_2    w  w_2    x  x_2    w  w_2    x  x_2
0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
1  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
2  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN
3  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN

Answer 1

您可以将join或merge与swaplevel()或reorder_levels 。 然后使用.sort_index()并通过axis=1按索引列排序。

• 当您像这样对索引进行合并时， .join()会更好。
• .swaplevel()在有两个级别时更好（如本例），而.reorder_levels()在 3 个或更多级别时更好。

以下是这些方法的 4 种组合。 对于这个特定的例子，我认为.join() / .swaplevel()是最疯狂的（见最后一个例子）：

df3 = (df1.reorder_levels([1,0],axis=1)
       .join(df2.reorder_levels([1,0],axis=1), rsuffix='_2')
       .reorder_levels([1,0],axis=1).sort_index(axis=1, level=[0, 1]))
df3
Out[1]: 
     a                   b                   c               
     w  w_2    x  x_2    w  w_2    x  x_2    w  w_2    x  x_2
0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
1  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
2  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN
3  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN

df3 = (pd.merge(df1.reorder_levels([1,0],axis=1),
                df2.reorder_levels([1,0],axis=1),
                how='left', left_index=True, right_index=True, suffixes = ('', '_2'))
                .reorder_levels([1,0],axis=1).sort_index(axis=1, level=[0, 1]))
df3
Out[2]: 
     a                   b                   c               
     w  w_2    x  x_2    w  w_2    x  x_2    w  w_2    x  x_2
0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
1  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
2  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN
3  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN

df3 = (pd.merge(df1.swaplevel(axis=1),
                df2.swaplevel(axis=1),
                how='left', left_index=True, right_index=True, suffixes = ('', '_2'))
                .swaplevel(axis=1).sort_index(axis=1, level=[0, 1]))
df3
Out[3]: 
     a                   b                   c               
     w  w_2    x  x_2    w  w_2    x  x_2    w  w_2    x  x_2
0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
1  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
2  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN
3  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN

df3 = (df1.swaplevel(i=0,j=1, axis=1)
       .join(df2.swaplevel(axis=1), rsuffix='_2')
       .swaplevel(axis=1).sort_index(axis=1, level=[0, 1]))
df3
Out[4]: 
     a                   b                   c               
     w  w_2    x  x_2    w  w_2    x  x_2    w  w_2    x  x_2
0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
1  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0  1.0  0.0
2  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN
3  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN  1.0  NaN