Applying Custom Functions (apply, map, applymap)0%
Vectorized String Methods (.str accessor)

Applying Custom Functions (apply, map, applymap)

Beginner12 min read•Updated: Jul 11, 2026
Study MaterialsCheatsheet

Applying Custom Functions (apply, map, applymap)

Reading Time: 8 Mins | Level: Beginner to Advanced

1. When Vectorization Isn't Enough

Pandas built-in vectorized operations are always the fastest approach. However, real-world business logic often requires custom branching (e.g. tax calculations, complex regex transformations, or calling external APIs). For these tasks, Pandas provides:

  • Series.map(): Map values of a Series using a dictionary or 1-to-1 function.
  • Series.apply() / DataFrame.apply(): Apply arbitrary functions across a Series or along DataFrame axes.
  • DataFrame.map() (formerly applymap in older Pandas): Apply a function element-wise across the entire table.

2. Using Series.map() for Value Substitution

Python
import pandas as pd
 
df = pd.DataFrame({
'Employee': ['Aarav', 'Priya', 'Rohan'],
'Gender_Code': ['M', 'F', 'M'],
'Grade': ['A', 'B', 'A']
})
 
# Dictionary mapping for quick translation
gender_map = {'M': 'Male', 'F': 'Female'}
df['Gender'] = df['Gender_Code'].map(gender_map)
print(df)

3. Using Series.apply() with Custom Functions & Lambdas

Python
def calculate_tax(salary):
if salary > 100000:
return salary * 0.30
elif salary > 50000:
return salary * 0.20
else:
return salary * 0.05
 
df_salaries = pd.DataFrame({'Emp': ['A', 'B', 'C'], 'Salary': [120000, 65000, 40000]})
 
# Apply custom function to Salary column
df_salaries['Tax_INR'] = df_salaries['Salary'].apply(calculate_tax)
 
# Or with an inline lambda function
df_salaries['Take_Home'] = df_salaries.apply(lambda row: row['Salary'] - row['Tax_INR'], axis=1)
print(df_salaries)
Pro Tip
When applying a function across multiple columns of a single row, pass axis=1 so the lambda receives the row as a Series!

4. Performance Warning: The Cost of apply()

Unlike native vectorized methods (which run compiled in C), .apply() is effectively a Python-level loop under the hood.

Always prefer native vectorized math where possible:

  • Fast: df['Total'] = df['Price'] * df['Qty'] (Vectorized C speed)
  • Slow: df['Total'] = df.apply(lambda r: r['Price'] * r['Qty'], axis=1) (Python loop speed)
Interactive Quiz

Multiple Choice Questions

Question 1
Which method is best suited for replacing values in a Series using a dictionary lookup table?
Question 2
When calling df.apply(custom_func, axis=1), what does the function receive as its argument?
Question 3
In Pandas 2.1+, which method replaces the deprecated DataFrame.applymap() for applying a function element-wise to every cell in a DataFrame?
Question 4
Why should you avoid using df.apply(..., axis=1) for simple arithmetic operations like Col_A + Col_B?
Question 5
What will pd.Series(['delhi', 'mumbai']).apply(str.upper) return?
Next Lesson

Vectorized String Methods (.str accessor)

Continue learning with hands-on practice, examples, and exercises in the upcoming topic.

Related Lessons

Previous LessonNext Lesson
Handling Duplicates & Data Type ConversionsVectorized String Methods (.str accessor)

Practice Quiz

Test your understanding of this lesson with 5 questions. Each question has one correct answer.

PreviousNext
PrevNext