如何修改 pandas 数据框多索引的 0 级？

Question

我有以下 pandas 数据框：

                                             0   value
name                    ga:browserVersion                  
Chrome 44.18            43.0.2357.130      0.139987   14.0%
                        43.0.2357.124      0.113107  11.31%
                        43.0.2357.134      0.103564  10.36%
                        44.0.2403.155      0.093181   9.32%
                        43.0.2357.81       0.092643   9.26%
                        44.0.2403.157      0.082780   8.28%
                        44.0.2403.125      0.070978    7.1%
                        44.0.2403.130      0.066152   6.62%
                        43.0.2357.132      0.064872   6.49%
                        44.0.2403.107      0.039940   3.99%
Internet Explorer 32.12 11.0               0.769828  76.98%
                        9.0                0.101842  10.18%
                        10.0               0.063672   6.37%
                        8.0                0.057929   5.79%
                        7.0                0.006320   0.63%
                        6.0                0.000353   0.04%
                        7.0b               0.000024    0.0%
                        999.1              0.000024    0.0%
                        10.6               0.000003    0.0%
                        5.5                0.000003    0.0%
Firefox 12.76           39.0               0.404164  40.42%
                        38.0               0.340139  34.01%
                        40.0               0.139032   13.9%
                        31.0               0.043926   4.39%
                        37.0               0.012160   1.22%
                        36.0               0.006963    0.7%
                        34.0               0.005601   0.56%
                        35.0               0.005495   0.55%
                        21.0               0.003508   0.35%
                        33.0               0.003209   0.32%
Safari 9.37             8.0.6              0.174829  17.48%
                        8.0.7              0.172087  17.21%
                        7.1.6              0.077686   7.77%
                        5.1.9              0.072729   7.27%
                        6.1.6              0.067831   6.78%
                        7.1.7              0.053092   5.31%
                        8.0.5              0.052637   5.26%
                        8.0.3              0.035921   3.59%
                        8.0.8              0.030222   3.02%
                        8.0.4              0.027923   2.79%
Opera 0.56              30.0.1835.125      0.220076  22.01%
                        30.0.1835.88       0.163912  16.39%
                        30.0.1835.59       0.123083  12.31%
                        31.0.1889.99       0.114718  11.47%
                        31.0.1889.174      0.111532  11.15%
                        29.0.1795.60       0.072296   7.23%
                        12.17              0.063334   6.33%
                        12.16              0.019319   1.93%
                        30.0.1835.52       0.009162   0.92%
                        29.0.1795.54600    0.008763   0.88%

这有一个包含 2 个级别的多索引，name 和 ga:browserVersion。我想要做的是将“%”添加到级别 0，使其看起来像：

Chrome 44.18%

等等

我创建了一个列表，其中包含我想用来替换当前索引的值：

new_index = []
for i in df.index.get_level_values(0):
    i = i+'%'
    new_index.append(i)

然后我尝试替换旧索引：

 df.index.get_level_values(0) = new_index

但我得到：

SyntaxError: can't assign to function call

我知道这对 'normal' 索引仅在一个级别上有效。有没有办法用多索引来实现这个？

Answer 1

您可以使用 rename 并传递新列名的字典，例如：

In [38]:
df = pd.DataFrame({'A' : ['foo', 'bar', 'foo', 'bar',
                           'foo', 'bar', 'foo', 'foo'],
                    'B' : ['one', 'one', 'two', 'three',
                           'two', 'two', 'one', 'three'],
                    'C' : np.random.randn(8), 'D' : np.random.randn(8)})
df

Out[38]:
     A      B         C         D
0  foo    one -0.510057  0.063085
1  bar    one -0.570631 -0.648810
2  foo    two -1.360048  1.609831
3  bar  three  0.628927 -0.379887
4  foo    two -0.415176 -1.798492
5  bar    two -0.147208 -0.366342
6  foo    one -0.333823  1.136703
7  foo  three  1.054773 -0.781997

In [36]:
mi = df.set_index(['A','B'])
mi

Out[36]:
                  C         D
A   B                        
foo one    0.172031  0.371076
bar one    1.007468  0.993607
foo two    0.552025 -0.478913
bar three  0.128154 -0.709580
foo two   -0.211721  0.569326
bar two   -0.713624  0.745678
foo one   -0.109175  0.448490
    three -0.388360  0.762513

In [39]:    
mi.rename(index={'foo':'yes','bar':'no'})

Out[39]:
                  C         D
A   B                        
yes one    0.172031  0.371076
no  one    1.007468  0.993607
yes two    0.552025 -0.478913
no  three  0.128154 -0.709580
yes two   -0.211721  0.569326
no  two   -0.713624  0.745678
yes one   -0.109175  0.448490
    three -0.388360  0.762513

所以在你的情况下，我将使用 zip:

创建一个字典

df.rename(index=dict(zip(df.index.get_level_values(0), new_index)))

Answer 2

# first reset the index , the indices should be converted to columns
df_reset = df.reset_index()

# then concatenate `%` to your column ( this column was an index level 0 )    
df_reset.name = df_reset.name + '%'

# set the index again for your data frame
df_reset.set_index(['name' , 'ga:browserVersion'])

如何修改 pandas 数据框多索引的 0 级？

How to modify a the level 0 of a pandas data frame multi index?

python

multi-index

pandas