如何对同一数据帧进行求和、平均、分组计数和标准差？

Question

我正在尝试按“名称”和“站点”对这个数据框进行分组，并且我想创建 4 个新列来查找“支出”列的总和、计数分组、平均值和标准差。

到目前为止，这是我的代码：

import pandas as pd

df=pd.DataFrame({'Name':['Harry','John','Holly','John','John','John','Holly','Holly','Molly','Molly','Holly','Harry','Harry','Harry'], 'Spend': [76,43,23,43,234,54,34,12,43,54,65,23,12,32],
                 'Site': ['Amazon','Ikea','Apple','Amazon', 'Apple', 'Ikea', 'Apple', 'Apple', 'Amazon', 'Amazon', 'Ikea', 'Amazon', 'Amazon', 'Ikea']})

print (df)

目前我的数据框如下所示：

我希望它看起来像这样：

我该怎么做呢？

提前致谢

编辑 2018 年 10 月 11 日：

代码：

import pandas as pd

df=pd.DataFrame({'Name':['Harry','John','Holly','John','John','John','Holly','Holly','Molly','Molly','Holly','Harry','Harry','Harry'], 'Spend': [76,43,23,43,234,54,34,12,43,54,65,23,12,32],
                 'Site': ['Amazon','Ikea','Apple','Amazon', 'Apple', 'Ikea', 'Apple', 'Apple', 'Amazon', 'Amazon', 'Ikea', 'Amazon', 'Amazon', 'Ikea'], 'Spend2': [176,143,123,143,1234,154,134,112,143,254,365,423,512,632]})

print (df)

前：

后：

Answer 1

df_summary = df.groupby(['Name', 'Site']).agg([np.sum, pd.Series.count, np.mean, np.std])
df_summary.columns = ['Sum', 'Count Groupbys', 'Average', 'Standard Deviation']
df_summary = df_summary.reset_index().sort_values(['Site', 'Name'])

>>> df_summary
    Name    Site  Sum  Count Groupbys  Average  Standard Deviation
0  Harry  Amazon  111               3     37.0           34.219877
4   John  Amazon   43               1     43.0                 NaN
7  Molly  Amazon   97               2     48.5            7.778175
2  Holly   Apple   69               3     23.0           11.000000
5   John   Apple  234               1    234.0                 NaN
1  Harry    Ikea   32               1     32.0                 NaN
3  Holly    Ikea   65               1     65.0                 NaN
6   John    Ikea   97               2     48.5            7.778175

根据您的编辑，您可以agg通过传递在列上键入的字典来使用，其中的值是应用于这些列的函数：

df_summary = df.groupby(['Name', 'Site']).agg(
    {'Spend': [np.sum, pd.Series.count], 
     'Spend2': [np.mean, np.std]}
)
df_summary.columns = ['Sum_Spend', 'CountGroupbys_Spend', 'Average_Spend2', 'Standard_Deviation_Spend2']
df_summary = df_summary.reset_index().sort_values(['Site', 'Name'])

>>> df_summary

    Name    Site    Sum_Spend   CountGroupbys_Spend Average_Spend2  Standard_Deviation_Spend2
0   Harry   Amazon  111        3    370.333333      174.081399
4   John    Amazon  43         1    143.000000      NaN
7   Molly   Amazon  97         2    198.500000      78.488853
2   Holly   Apple   69         3    123.000000      11.000000
5   John    Apple   234        1    1234.000000     NaN
1   Harry   Ikea    32         1    632.000000      NaN
3   Holly   Ikea    65         1    365.000000      NaN
6   John    Ikea    97         2    148.500000      7.778175