根据 A 列的值对 B 列的值应用操作 - Pandas

Question

I have the following dataset:

id  Wages   PayFreq
0   1013    Weekly
1   5000    Monthly
2   892     Weekly
3   2320    Bi-Weekly
4   1068    Weekly

I intend to perform the following operation:

if PayFreq == 'Monthly':
   (Wages / 4) * 52
elif PayFreq == 'Bi-Weekly':
   (Wages / 2) * 52
else:
    Wages * 52

I need to select the operation to apply to the wages column based on what is present in the PayFreq column. Any ideas?

Answer 1

Use Series.map by dictionary, for not matched values use Series.fillna because map return NaN and multiple by Wages column:

d = {'Monthly': 52/4, 'Bi-Weekly': 52/2}
df['YearWages'] = df['PayFreq'].map(d).fillna(52).mul(df['Wages'])

print (df)
   id  Wages    PayFreq  YearWages
0   0   1013     Weekly    52676.0
1   1   5000    Monthly    65000.0
2   2    892     Weekly    46384.0
3   3   2320  Bi-Weekly    60320.0
4   4   1068     Weekly    55536.0

Solution with masks passed to numpy.select :

df['NewWages'] = np.select([df['PayFreq'] == 'Weekly',
                            df['PayFreq']== ' Monthly'], 
                            [(df['Wages'] / 2)*52, 
                             (df['Wages'] / 4) * 52], default=df['Wages']*52)
                            
print (df)
   id  Wages    PayFreq  NewWages
0   0   1013     Weekly   26338.0
1   1   5000    Monthly  260000.0
2   2    892     Weekly   23192.0
3   3   2320  Bi-Weekly  120640.0
4   4   1068     Weekly   27768.0

Answer 2

Or you can use np.where , and adjust the conditions accordingly, use the below just as an example not as a full answer :

import numpy as np
df['NewWages'] = np.where(df['PayFreq'] == 'Weekly', (df['Wages'] / 2)*52,
                          np.where(df['PayFreq']== ' Monthly', (df['Wages'] / 4) * 52, df['Wages']*52))

which prints:

   id  Wages    PayFreq  NewWages
0   0   1013     Weekly   26338.0
1   1   5000    Monthly  260000.0
2   2    892     Weekly   23192.0
3   3   2320  Bi-Weekly  120640.0
4   4   1068     Weekly   27768.0

Answer 3

I would sugest to use Dataframe's apply as it is very simple and intuitive.

You can define a method you want to apply on the dataframe, you can choose either Lambda expression or explicit function for that. For example, here is a simple implementation with a function:

def func(row):
    if row['PayFreq'] == 'Monthly':
        return (row['Wages'] / 4) * 52
    elif row['PayFreq'] == 'Bi-Weekly':
        return (row['Wages'] / 2) * 52
    else:
        return row['Wages'] * 52

And in order to apply it on your dataframe (on the right axis):

df['NewWages'] = df.apply(func, axis=1)

The result:

   Wages    PayFreq  NewWages
0   1013     Weekly   52676.0
1   5000    Monthly   65000.0
2    892     Weekly   46384.0
3   2320  Bi-Weekly   60320.0
4   1068     Weekly   55536.0