基于多列排序的RANK

Question

我正在尝试按 ID_1 分组并按 ID_2 降序和 TotalRevenue 升序排序以提取排名

请协助如何在RANK排序函数中同时使用升序和降序功能

import pandas as pd

df = pd.DataFrame({

    'ID_1':[1,1,1,2,2,2,3,3],
    'ID_2':[100,100,35,30,30,20,50,50],
    'TotalRevenue':[9000,2000,750,1000,600,500,500,300]})
  

df['RANK']= df.groupby(['ID_1'])['ID_2','TotalRevenue'].rank(method='dense',ascending=False).astype(int)


Desired output:

ID_1   ID_2   TotalRevenue       Rank
   1   100          9000           2
   1   100          2000           1
   1   35           750            3
   2   30           1000           2
   2   30           600            1
   2   20           500            3
   3   50           500            2
   3   50           300            1

Answer 1

你可以sort_values并根据组分配cumcount+1

out = df.sort_values(['ID_2','TotalRevenue'],ascending=[False,True])
out['Rank'] = out.groupby("ID_1").cumcount()+1
print(out.sort_index())

   ID_1  ID_2  TotalRevenue  Rank
0     1   100          9000     2
1     1   100          2000     1
2     1    35           750     3
3     2    30          1000     2
4     2    30           600     1
5     2    20           500     3
6     3    50           500     2
7     3    50           300     1

基于多列排序的RANK

RANK based on SORT by multiple columns

python

sorting

numpy

rank

pandas