从DataFrame中删除NaN，并从多索引中删除重复项

问题描述

给出一个多索引DataFrame:

Given a multi-index DataFrame:

                 col_A  col_B
level_0 level_1              
A       x          1.0    NaN
        y          NaN    1.0
        x          NaN    2.0
        y          2.0    NaN

如何从df中删除NaN，并从多索引中删除重复项以获取:

How can I remove the NaNs from the df and duplicates from the multi-index to get:

                 col_A  col_B
level_0 level_1              
A       x          1.0    2.0
        y          2.0    1.0

这是MWE:

import pandas as pd
import numpy as np

index = pd.MultiIndex.from_product([['A', 'A'],
                                  ['x', 'y']],
                                 names=['level_0',
                                        'level_1'])
data =[
    [1, np.NaN],
    [np.NaN, 1],
    [np.NaN,2],
    [2, np.NaN],
]
df = pd.DataFrame(data=data, index=index, columns=['col_A', 'col_B'])
print df

推荐答案

在index名称上使用groupby，并获取first值.

Use groupby on index names, and take first values.

In [642]: df.groupby(level=df.index.names).first()
Out[642]:
                 col_A  col_B
level_0 level_1
A       x          1.0    2.0
        y          2.0    1.0

注:编辑后，意识到它几乎与Psidom的答案相同.对level

Note: Post edit, realized it's almost identical to Psidom's answer. A minor generic edit to level

这篇关于从DataFrame中删除NaN，并从多索引中删除重复项的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！