数据框分组依据-行到列

Question

DF有

,manufacturer,project,type,Metric
0,Honda,project_a,sedan,10
1,Honda,project_a,suv,20
2,Honda,project_a,hatchback,2
3,Toyota,project_a,sedan,11
4,Toyota,project_a,suv,21
5,Toyota,project_a,hatchback,3
6,Honda,project_b,sedan,101
7,Honda,project_b,suv,201
8,Honda,project_b,hatchback,21
9,Toyota,project_b,sedan,111
10,Toyota,project_b,suv,211
11,Toyota,project_b,hatchback,31

DF想要

                     project_a                   project_b
type          hatchback  sedan  suv          hatchback  sedan  suv
manufacturer                       
Honda                 2     10   20 .        21     101   201
Toyota                3     11   21 .        31     111   211

目前我有-

df1 ['Metric']。groupby（[df1 ['manufacturer']，df1 ['type']]）。mean（）。unstack（）

type          hatchback  sedan    suv
manufacturer                         
Honda              11.5   55.5  110.5
Toyota             17.0   61.0  116.0

如何将project_a和project_b扩展为更高级别的列？

Answer 1

好像您想pivot_table多列

df.pivot_table(columns=['project', 'type'], index=['manufacturer'], values=['Metric'])

project         project_a                      project_b
type            hatchback   sedan   suv        hatchback    sedan   suv
manufacturer                        
Honda           2           10      20         21           101     201
Toyota          3           11      21         31           111     211