将平均列添加到Pandas Multiindex数据框

Question

我有一个数据框df

first        bar                 baz           
second       one       two       one       two 
A       0.487880 -0.487661 -1.030176  0.100813 
B       0.267913  1.918923  0.132791  0.178503
C       1.550526 -0.312235 -1.177689 -0.081596 

我想添加平均值列，然后将平均值移到最前面

df['Average'] = df.mean(level='second', axis='columns')  #ERROR HERE
cols = df.columns.tolist()
df = df[[cols[-1]] + cols[:-1]]

我得到了错误：

ValueError: Wrong number of items passed 2, placement implies 1

也许我可以一次平均地将每一列df['Average', 'One'] = ...添加一次，但这似乎很愚蠢，尤其是当现实生活中的索引更加复杂时。

编辑：（Frame Generation）

arrays = [['bar', 'bar', 'baz', 'baz', 'foo', 'foo', 'qux', 'qux'],
    ['one', 'two', 'one', 'two', 'one', 'two', 'one', 'two']]

tuples = list(zip(*arrays))
index = pd.MultiIndex.from_tuples(tuples, names=['first', 'second'])
df = DataFrame(np.random.randn(3, 8), index=['A', 'B', 'C'], columns=index)

Answer 1

我不确定您的目标输出。像这样吗？

df2 = df.mean(level='second', axis='columns')
df2.columns = pd.MultiIndex.from_tuples([('mean', col) for col in df2])
>>> df2
       mean          
        one       two
A -0.271148 -0.193424
B  0.200352  1.048713
C  0.186419 -0.196915

>>> pd.concat([df2, df], axis=1)
       mean                 bar                 baz          
        one       two       one       two       one       two
A -0.271148 -0.193424  0.487880 -0.487661 -1.030176  0.100813
B  0.200352  1.048713  0.267913  1.918923  0.132791  0.178503
C  0.186419 -0.196915  1.550526 -0.312235 -1.177689 -0.081596

由于mean操作会导致一个数据帧（在这种情况下为两列），因此会出现错误。然后，您尝试将此结果分配到原始数据帧的一栏中。

Answer 2

pandas.concat

df.join(pd.concat([df.mean(level='second', axis='columns')], axis=1, keys=['Average']))

first        bar                 baz             Average          
second       one       two       one       two       one       two
A       0.255301  0.286846  1.027024 -0.060594  0.641162  0.113126
B      -0.608509 -2.291201  0.675753 -0.416156  0.033622 -1.353679
C       2.714254 -1.330621 -0.099545  0.616833  1.307354 -0.356894

---

stack / unstack

不一定高效，但是很整齐

df.stack().assign(Average=df.mean(level='second', axis='columns').stack()).unstack()

first        bar                 baz             Average          
second       one       two       one       two       one       two
A       0.255301  0.286846  1.027024 -0.060594  0.641162  0.113126
B      -0.608509 -2.291201  0.675753 -0.416156  0.033622 -1.353679
C       2.714254 -1.330621 -0.099545  0.616833  1.307354 -0.356894