将Pandas列乘以多索引值

Question

我有一个包含三个索引级别的多索引数据框。 我想将每个“百分比”值乘以总人口。

我正在尝试这样的事情：

df = df.reset_index(level=2, drop=True)
df['TOTAL'] = df['POP.'] * output['PERCENT']

我收到KeyError: 'POP.'

启动DataFrame

                               PERCENT  
     DATE      POP.  SEX                
2015-01-01     100  MALE        0.51  
                    FEMALE      0.49  
2016-01-01     200  MALE        0.52  
                    FEMALE      0.48

期望的输出

                               PERCENT     TOTAL  
     DATE      POP.  SEX                
2015-01-01     100  MALE        0.51         51  
                    FEMALE      0.49         49
2016-01-01     200  MALE        0.52        104
                    FEMALE      0.48         96

Answer 1

您需要get_level_values + values

df['TOTAL']=(df.index.get_level_values('POP')*df['PERCENT']).values
df
Out[874]: 
                       PERCENT  TOTAL
DATE       POP SEX                   
2015-01-01 100 MALE       0.51   51.0
               FEMALE     0.49   49.0
2016-01-01 120 Male       0.52   62.4
               FEMALE     0.48   57.6

Answer 2

您还可以使用multiply方法并按如下所示设置axis="index" ：

df['TOTAL'] = df['POP'].multiply(df['PERCENT'], axis="index")

Answer 3

您可以使用DataFrame.eval（）方法：

In [118]: df = df.eval("TOTAL = PERCENT * POP", inplace=False)

In [119]: df
Out[119]:
                         PERCENT  TOTAL
DATE       POP   SEX
2015-01-01 100.0 MALE       0.51   51.0
                 FEMALE     0.49   49.0
2016-01-01 200.0 MALE       0.52  104.0
                 FEMALE     0.48   96.0