.assign()

MamtaWardhani's avatar
MamtaWardhani's avatar
MamtaWardhani
Published Jul 2, 2025
Contribute to Docs

The .assign() method in pandas is used to create new columns or modify existing columns in a DataFrame while preserving the original DataFrame. It returns a new DataFrame object with all original columns plus the newly assigned ones, making it an ideal choice for functional programming approaches and method chaining.

The .assign() method is particularly useful in data preprocessing, feature engineering, and exploratory data analysis workflows. It allows for dynamic column creation using both static values and callable functions, making it versatile for various data manipulation scenarios. Common use cases include calculating derived metrics, creating categorical variables, applying transformations to existing columns, and building complex data pipelines where immutability is preferred.

Syntax

DataFrame.assign(**kwargs)

Parameters:

  • **kwargs: Keyword arguments where the keys represent column names and values can be:
    • Scalar values: Single values assigned to all rows
    • Array-like objects: Lists, arrays, or Series with length matching the DataFrame
    • Callable functions: Functions or lambda expressions that take the DataFrame as input and return values for the new column

Return value:

A new DataFrame containing all original columns plus the newly assigned columns. If a column name already exists, it will be replaced in the returned DataFrame.

Example 1: Adding Basic Columns

This example demonstrates the fundamental usage of .assign() to add new columns with static values:

import pandas as pd


# Create sample DataFrame
df = pd.DataFrame({
  'product_id': [101, 102, 103, 104],
  'price': [25.99, 45.50, 12.75, 67.25],
  'quantity': [10, 5, 20, 8]
})


# Add new columns using .assign()
df_updated = df.assign(
  currency='USD',
  in_stock=True,
  category='Electronics'
)


print(df_updated)

Output of this code is:

   product_id  price  quantity currency  in_stock    category
0         101  25.99        10      USD      True  Electronics
1         102  45.50         5      USD      True  Electronics
2         103  12.75        20      USD      True  Electronics
3         104  67.25         8      USD      True  Electronics

This example shows how .assign() creates a new DataFrame with three additional columns. The currency and category columns receive string values, while in_stock receives a boolean value that is applied to all rows.

Example 2: Calculating Revenue Analysis

This example demonstrates using .assign() with lambda functions to perform business calculations and create derived metrics:

import pandas as pd


# Create sales DataFrame
sales_df = pd.DataFrame({
  'product_name': ['Laptop', 'Mouse', 'Keyboard', 'Monitor'],
  'unit_price': [899.99, 29.99, 79.99, 299.99],
  'units_sold': [50, 200, 150, 75],
  'cost_per_unit': [600.00, 15.00, 45.00, 180.00]
})


# Calculate business metrics using .assign()
sales_analysis = sales_df.assign(
  total_revenue=lambda x: x['unit_price'] * x['units_sold'],
  total_cost=lambda x: x['cost_per_unit'] * x['units_sold'],
  profit=lambda x: (x['unit_price'] - x['cost_per_unit']) * x['units_sold'],
  profit_margin=lambda x: ((x['unit_price'] - x['cost_per_unit']) / x['unit_price']) * 100
)


print(sales_analysis.round(2))

The output of this code is:

  product_name  unit_price  units_sold  cost_per_unit  total_revenue  total_cost   profit  profit_margin
0       Laptop      899.99          50         600.00       44999.50     30000.0  14999.5          33.33
1        Mouse       29.99         200          15.00        5998.00      3000.0   2998.0          50.02
2     Keyboard       79.99         150          45.00       11998.50      6750.0   5248.5          43.76
3      Monitor      299.99          75         180.00       22499.25     13500.0   8999.25         40.00

This example demonstrates real-world financial analysis where .assign() calculates multiple business metrics. Each lambda function receives the DataFrame as input and performs calculations using existing columns to create comprehensive sales analytics.

Codebyte Example: Customer Segmentation Pipeline

This example shows advanced usage of .assign() for customer segmentation and data categorization in a chained workflow:

Code
Output
Loading...


This advanced example demonstrates how .assign() can be used for complex customer segmentation. It creates multiple categorical variables using pd.cut(), conditional logic with np.where(), and mathematical calculations, all within a single method call to build a comprehensive customer analysis framework.


Frequently Asked Questions


1. Does .assign() modify the original DataFrame?


No, .assign() returns a new DataFrame with the assigned columns while leaving the original DataFrame unchanged.


2. Can I assign multiple columns in a single .assign() call?


Yes, you can assign multiple columns by providing multiple keyword arguments. Later columns in the same .assign() call can reference earlier columns that were created in the same call.


3. How does .assign() handle existing column names?


If you assign to an existing column name, .assign() will overwrite that column in the returned DataFrame. The original DataFrame remains unchanged, but the new DataFrame will have the updated column values.


All contributors
Contribute to Docs
Learn Python:Pandas on Codecademy