将一列添加到对一列求和的多索引中

Question

我想添加另一列，其中包含每个本国的工资列的总和，并插入每个本国的薪资列&工资组合。

这就是我尝试过的

a = df.groupby(['native-country','salary'])[['salary']].count()
a.columns = ['number']
a['total'] = a.sum(level=0, axis=0)

哪个返回

number  total
native-country salary               
?              <=50K      437    NaN
>50K       146    NaN
Cambodia       <=50K       12    NaN
>50K         7    NaN
Canada         <=50K       82    NaN
...                       ...    ...
United-States  >50K      7171    NaN
Vietnam        <=50K       62    NaN
>50K         5    NaN
Yugoslavia     <=50K       10    NaN
>50K         6    NaN
[82 rows x 2 columns]

Answer 1

根据您的评论，我认为这就是您想要的。但你的问题仍然不清楚，所以也许我的假设是错误的。

a['total'] = a.groupby(level=0)["Number].sum()

Answer 2

试试这个：

a = df.groupby(['native-country','salary'])[['salary']].count()
a.join(a.sum(level=0).rename(columns={'salary':'Total'}))

这是一个MVCE:

import seaborn as sns
import pandas as pd
df_tips = sns.load_dataset('tips')
df_out = df_tips.groupby(['sex','smoker'])[['tip']].sum()
df_out.join(df_out.sum(level=0).rename(columns={'tip':'total'}))

输出：

tip   total
sex    smoker                
Male   Yes     183.07  485.07
No      302.00  485.07
Female Yes      96.74  246.51
No      149.77  246.51