[FIXED] Second run of 'np.where' overwrites changes from first run

Issue

I am trying to add a new column in pandas with some text if the data in the 'Level' column is found in a list. It works; however, the last np.where is overwriting the changes made by the first. I am not sure how to get the all the changes to stay. I realize that the second np.where is simply overwriting the changes made by the first, but cannot figure out how to apply both of them. I tried a for loop, an if, and making a copy... newdf = ppl.copy(), but then was not sure how to move the contents into the copy. Digging around I think I might need ppl.loc? Im sure I am missing something simple.

---

import pandas as pd
import numpy as np
ppl = pd.read_excel('MyXLS.xlsx', sheet_name='People')
list8 = ['8-01','8-02','8-03','8-59']
list9 = ['9-01','9-04','9-36']
LVL8 = "These change Tuesday"
LVL9 = "These change Saturday" 

ppl['Weekday'] = np.where(ppl['Level'].isin(list8), LVL8, 'NA') 
print(ppl.head(10))
ppl['Weekday'] = np.where(ppl['Level'].isin(list9), LVL9, 'NA') 
print(ppl.head(10))

The ouput:

     Name Level               Weekday
0    Mark  8-01  These change Tuesday
1    Gary  8-02  These change Tuesday
2    Lisa  8-03  These change Tuesday
3    John  8-59  These change Tuesday
4  Jessie  9-01                    NA
5   Chris  9-04                    NA
6     Sam  9-36                    NA
     Name Level                Weekday
0    Mark  8-01                     NA
1    Gary  8-02                     NA
2    Lisa  8-03                     NA
3    John  8-59                     NA
4  Jessie  9-01  These change Saturday
5   Chris  9-04  These change Saturday
6     Sam  9-36  These change Saturday

Solution

You could nest your np.where calls so that the second doesn't overwrite the changes made by the first:

ppl['Weekday'] = np.where(ppl['Level'].isin(list9), LVL9, np.where(ppl['Level'].isin(list8), LVL8, 'NA'))

But it's simpler to just use .loc to address the specific rows:

ppl['Weekday'] = 'NA'
ppl.loc[ppl['Level'].isin(list8), 'Weekday'] = LVL8
ppl.loc[ppl['Level'].isin(list9), 'Weekday'] = LVL9

or perhaps np.select:

ppl['Weekday'] = np.select([ ppl['Level'].isin(list8), ppl['Level'].isin(list9) ], [ LVL8, LVL9 ], 'NA')

Answered By - Nick