Groupby在Python中的多个条件下求和多个列

Question

我有一个像这样的pandas数据框:

import pandas as pd
import numpy as np
data = {
    "Type": ["A", "A", "B", "B", "B"],
    "Project": ["X123", "X123", "X21", "L31", "L31"],
    "Number": [100, 300, 100, 200, 500],
    "Status": ['Y', 'Y', 'N', 'Y', 'N']
}
df = pd.DataFrame.from_dict(data)

我想按类型进行分组,并获得计数和几个条件的总和,得到如下结果:

Type  Total_Count  Total_Number  Count_Status=Y  Number_Status=Y  Count_Status=N  Number_Status=N 
 A        2          400              2               400              0               0
 B        5          800              1               200              2              600

我试过以下但不完全是我需要的.请分享您可能有的任何想法.谢谢!

df1 = pd.pivot_table(df, index = 'Type', values = 'Number', aggfunc = np.sum)
df2 = pd.pivot_table(df, index = 'Type', values = 'Project', aggfunc = 'count')
pd.concat([df1, df2], axis=1)

Answer 1

如果要创建函数:

def my_agg(x):
    names = {
        'Total_Count': x['Type'].count(),
        'Total_Number': x['Number'].sum(),
        'Count_Status=Y': x[x['Status']=='Y']['Type'].count(),
        'Number_Status=Y': x[x['Status']=='Y']['Number'].sum(),
        'Count_Status=N': x[x['Status']=='N']['Type'].count(),
        'Number_Status=N': x[x['Status']=='N']['Number'].sum()}

    return pd.Series(names)

df.groupby('Type').apply(my_agg)

    Total_Count   Total_Number  Count_Status=Y  Number_Status=Y Count_Status=N  Number_Status=N
Type                        
A      2           400                2                400            0             0
B      3           800                1                200            2            600