Question

在pandas版本0.19.2上，我有以下带有multiindex的数据框：

import pandas as pd
import numpy as np

arrays = [[2001, 2001, 2002,  2002, 2002, 2003, 2004, 2004],
          ['A', 'B', 'A', 'C', 'D', 'B', 'C', 'D']]
tuples = list(zip(*arrays))
index = pd.MultiIndex.from_tuples(tuples, names=['first', 'second'])
s = pd.Series(np.random.randn(8), index=index, name='signal')

如下所示：

first  second
2001   A        -2.48
       B         0.95
2002   A         0.55
       C         0.65
       D        -1.32
2003   B        -0.25
2004   C         0.86
       D        -0.31

我希望获得一个摘要列联数据框，其中列是唯一的＆＃34;第二个＆＃34;和指数是＆＃34;第一＆＃34;索引，如下所示：

       A      B     C     D
2001  -2.48  0.95   NaN   NaN
2002  0.55   NaN    0.65
2003  NaN    -0.25  NaN   NaN
2004  NaN    NaN    0.86  -0.31

知道如何做到这一点？我玩groupby()如下所示但无法到达任何地方

s.groupby(level=1).apply(lambda x: "to do")

关联问题：Python Pandas - how to do group by on a multiindex 如何-待办事项基团的由式-A-多指标

Answer 1

如果MultiIndex中的唯一对：

，我认为需要unstack

df = s.unstack()
print (df)
second         A         B         C         D
first                                         
2001    1.752237  0.348548       NaN       NaN
2002   -0.022903       NaN -0.961702  0.079236
2003         NaN -0.393272       NaN       NaN
2004         NaN       NaN -0.600994 -0.594842

但如果在实际数据中获得：

ValueError：索引包含重复的条目，无法重塑

这意味着MultiIndex中有重复项：

print (s)
first  second
2001   A         0.478052 <-2001, A
       A         0.485261 <-2001, A
2002   A        -0.474997
       C        -1.165866
       D        -0.755630
2003   B         0.588104
2004   C        -1.439245
       D        -0.461221
Name: signal, dtype: float64

然后可能的解决方案首先是聚合值：

print (s.groupby(level=[0,1]).mean())
first  second
2001   A         0.958668
2002   A        -0.459612
       C         0.534821
       D         1.469257
2003   B        -1.103208
2004   C         0.098037
       D         0.722135
Name: signal, dtype: float64

df = s.groupby(level=[0,1]).mean().unstack()
print (df)
second         A         B         C         D
first                                         
2001    0.481657       NaN       NaN       NaN
2002   -0.474997       NaN -1.165866 -0.755630
2003         NaN  0.588104       NaN       NaN
2004         NaN       NaN -1.439245 -0.461221

熊猫重新排列Multiindex系列

1 个答案: