Implement own function

April 18, 2022

I am trying to implement my own function with the data set below:

import pandas as pd
import numpy as np
    
data = { 
            'sales': ['0','1','2','2','6','5','6'],            
           }        
df = pd.DataFrame(data, columns = ['sales'])
    
df

Now I want to apply my function how will give 1 only on value ‘2’, 2 only on values ‘6’, and for all others will give ‘3’. In order to do this, I try this function :

def function_test(data):
    sales = df['sales']
    
    conditions = [
                 (sales == '6'), 
                 (sales == '2'),
                 (sales <> '6'&'2') #<----This row
                 
                 ]
    values = [ 
             1,
             2,
             3
             ]
    dummy = np.select(conditions, values)    
    return (dummy)

But this function has a problem for third conditions, so can anybody help me how to solve this problem?

>Solution :

One way to fix it is to use != instead of <> and use two comparisons. (I also changed the condition sequence to match what your described in the text of your question):

((sales != '6') & (sales != '2'))

Full test code:

def function_test(data):
    sales = df['sales']
    conditions = [
                 (sales == '2'),
                 (sales == '6'), 
                 ((sales != '6') & (sales != '2'))
                 ]
    values = [1, 2, 3]
    dummy = np.select(conditions, values)    
    return (dummy)

print(function_test(data))

Results: