How to Replace Values in Columns Based on Condition in Pandas

In Python, we can replace values in columns based on conditions in Pandas using various built-in functions like loc, where and mask, apply and lambda, map, and numpy.where. Pandas is a powerful library for data manipulation and working with structured data. This article demonstrates five different methods to conditionally replace column values.

Using loc

The loc function allows you to access and modify specific rows and columns in a DataFrame based on boolean conditions ?

Syntax

df.loc[row_condition, column_labels] = new_value

Example

Let's replace the gender of people aged 50 or older with 'M' ?

import pandas as pd

data = {
    'name': ['Alice', 'Bob', 'Charlie', 'David', 'Emily'],
    'age': [25, 35, 45, 55, 65],
    'gender': ['F', 'M', 'M', 'F', 'F']
}

df = pd.DataFrame(data)
print("Original DataFrame:")
print(df)

# Replace gender with 'M' where age >= 50
df.loc[df['age'] >= 50, 'gender'] = 'M'
print("\nAfter replacement:")
print(df)

Original DataFrame:
      name  age gender
0    Alice   25      F
1      Bob   35      M
2  Charlie   45      M
3    David   55      F
4    Emily   65      F

After replacement:
      name  age gender
0    Alice   25      F
1      Bob   35      M
2  Charlie   45      M
3    David   55      M
4    Emily   65      M

Using where and mask

The where function keeps values where the condition is True and replaces others. The mask function does the opposite it replaces values where the condition is True ?

Using where()

import pandas as pd

data = {
    'name': ['Alice', 'Bob', 'Charlie', 'David', 'Emily'],
    'age': [25, 35, 45, 55, 65],
    'gender': ['F', 'M', 'M', 'F', 'F']
}

df = pd.DataFrame(data)

# Replace age with 0 where gender is 'M'
df['age'] = df['age'].where(df['gender'] != 'M', 0)
print(df)

      name  age gender
0    Alice   25      F
1      Bob    0      M
2  Charlie    0      M
3    David   55      F
4    Emily   65      F

Using mask()

import pandas as pd

data = {
    'name': ['Alice', 'Bob', 'Charlie', 'David', 'Emily'],
    'age': [25, 35, 45, 55, 65],
    'gender': ['F', 'M', 'M', 'F', 'F']
}

df = pd.DataFrame(data)

# Same result using mask - replace where condition is True
df['age'] = df['age'].mask(df['gender'] == 'M', 0)
print(df)

      name  age gender
0    Alice   25      F
1      Bob    0      M
2  Charlie    0      M
3    David   55      F
4    Emily   65      F

Using apply and Lambda

The apply function with lambda expressions provides flexible row-wise or column-wise transformations ?

import pandas as pd

data = {
    'name': ['Alice', 'Bob', 'Charlie', 'David', 'Emily'],
    'age': [25, 35, 45, 55, 65],
    'gender': ['F', 'M', 'M', 'F', 'F']
}

df = pd.DataFrame(data)

# Replace gender with 'F' where name starts with 'A'
df['gender'] = df.apply(lambda x: 'F' if x['name'].startswith('A') else x['gender'], axis=1)
print(df)

      name  age gender
0    Alice   25      F
1      Bob   35      M
2  Charlie   45      M
3    David   55      F
4    Emily   65      F

Using numpy.where()

NumPy's where function provides a vectorized approach for conditional replacement ?

import pandas as pd
import numpy as np

data = {
    'name': ['Alice', 'Bob', 'Charlie', 'David', 'Emily'],
    'age': [25, 35, 45, 55, 65],
    'gender': ['F', 'M', 'M', 'F', 'F']
}

df = pd.DataFrame(data)

# Replace age with 0 where gender is 'M', keep original age otherwise
df['age'] = np.where(df['gender'] == 'M', 0, df['age'])
print(df)

      name  age gender
0    Alice   25      F
1      Bob    0      M
2  Charlie    0      M
3    David   55      F
4    Emily   65      F

Comparison

Conclusion

Pandas offers multiple methods for conditional value replacement. Use loc for straightforward boolean indexing, numpy.where for best performance, and apply + lambda for complex logic. Choose the method that best fits your specific use case and performance requirements.