Appending a level (with fixed value) to pandas Series/DataFrame

Question

I have a pandas Series with Multi-index as follows:

category_1  number
A           0         1.764052
            1         0.400157
            2         0.978738
            3         2.240893
            4         1.867558
C           0        -0.977278
            1         0.950088
            2        -0.151357
            3        -0.103219
            4         0.410599

It is generated from this code:

import pandas as pd
import numpy as np
idx = pd.MultiIndex.from_product([['A','C'],range(5)], names=['category_1','number'])
np.random.seed(0)
s = pd.Series(index=idx, data = np.random.randn(len(idx)))

I would like to add another level, called category_2 to the index with a fixed value (ie D ) to have the following result:

category_1  category_2  number
A           D           0         1.764052
                        1         0.400157
                        2         0.978738
                        3         2.240893
                        4         1.867558
C           D           0        -0.977278
                        1         0.950088
                        2        -0.151357
                        3        -0.103219
                        4         0.410599

I have been using this hacky way to do this:

df =s.to_frame('dummy')
df['category_2'] = 'D'
df.set_index('category_2', append = True, inplace = True)
df = df.reorder_levels([0,2,1])
res = df['dummy']

Is there a better (more succinct/pythonic) way to add a level with fixed value to the existing levels on a pandas Series/DataFrame?

Answer 1

You need create new MultiIndex and then replace old one:

#change multiindex
new_index = list(zip(s.index.get_level_values('category_1'), 
                     ['D'] * len(s.index), 
                     s.index.get_level_values('number')))
print (new_index)
[('A', 'D', 0), ('A', 'D', 1),
 ('A', 'D', 2), ('A', 'D', 3), 
 ('A', 'D', 4), ('C', 'D', 0), 
 ('C', 'D', 1), ('C', 'D', 2), 
 ('C', 'D', 3), ('C', 'D', 4)]
s.index = pd.MultiIndex.from_tuples(new_index, 
                                    names=['category_1','category_2','number'])
print (s)
category_1  category_2  number
A           D           0         1.764052
                        1         0.400157
                        2         0.978738
                        3         2.240893
                        4         1.867558
C           D           0        -0.977278
                        1         0.950088
                        2        -0.151357
                        3        -0.103219
                        4         0.410599
dtype: float64

Another nice solution with MultiIndex.from_product - a bit changed comment :

s.index = pd.MultiIndex.from_product([s.index.levels[0], 
                                      ['D'], 
                                      s.index.levels[1]], names= ['c1','c2','number']) 
print (s)
c1  c2  number
A   D   0         1.764052
        1         0.400157
        2         0.978738
        3         2.240893
        4         1.867558
C   D   0        -0.977278
        1         0.950088
        2        -0.151357
        3        -0.103219
        4         0.410599
dtype: float64

Or:

s.index = pd.MultiIndex.from_product([s.index.get_level_values('category_1').unique(), 
                                      ['D'],  
                                      s.index.get_level_values('number').unique()], 
                                     names= ['c1','c2','number']) 
print (s)
c1  c2  number
A   D   0         1.764052
        1         0.400157
        2         0.978738
        3         2.240893
        4         1.867558
C   D   0        -0.977278
        1         0.950088
        2        -0.151357
        3        -0.103219
        4         0.410599
dtype: float64