In [1]:

    

import pandas as pd


    

In [2]:

    

classes = ["class 1"] * 5 + ["class 2"] * 5
sub_class = ['c1','c2','c2','c1','c3'] + ['c1','c2','c3','c2','c3']
vals = [1,3,5,1,3] + [2,6,7,5,2]
p_df = pd.DataFrame({"class": classes, "sub_class": sub_class, "vals": vals})


    

In [3]:

    

p_df


    

    
Out[3]:






  

    

      

      
class
      
sub_class
      
vals
    
  
  

    

      
0
      
class 1
      
c1
      
1
    
    

      
1
      
class 1
      
c2
      
3
    
    

      
2
      
class 1
      
c2
      
5
    
    

      
3
      
class 1
      
c1
      
1
    
    

      
4
      
class 1
      
c3
      
3
    
    

      
5
      
class 2
      
c1
      
2
    
    

      
6
      
class 2
      
c2
      
6
    
    

      
7
      
class 2
      
c3
      
7
    
    

      
8
      
class 2
      
c2
      
5
    
    

      
9
      
class 2
      
c3
      
2
    
  

In [4]:

    

grouped = p_df.groupby(['class', 'sub_class'])['vals'].median()


    

In [5]:

    

grouped


    

    
Out[5]:





class    sub_class
class 1  c1           1.0
         c2           4.0
         c3           3.0
class 2  c1           2.0
         c2           5.5
         c3           4.5
Name: vals, dtype: float64

In [6]:

    

for index_val, value in grouped.iteritems():
    class_name, sub_class_name = index_val
    print(class_name, ":", sub_class_name, ":", value)


    

    




class 1 : c1 : 1.0
class 1 : c2 : 4.0
class 1 : c3 : 3.0
class 2 : c1 : 2.0
class 2 : c2 : 5.5
class 2 : c3 : 4.5

In [ ]: