concat() for stacking DataFrames in Pandas

Introduction

We use concat() to join two or more tables (DataFrames) together. It helps us combine data from different sources into one big table.

You have sales data from January and February in separate tables and want to see all months together.

You collected survey results from different groups and want to combine them into one list.

You want to add new rows of data to an existing table without changing the columns.

You want to stack data vertically or horizontally to compare or analyze together.

Syntax

Pandas

import pandas as pd

# Concatenate DataFrames vertically (stack rows)
combined_df = pd.concat([df1, df2], axis=0, ignore_index=True)

# Concatenate DataFrames horizontally (stack columns)
combined_df = pd.concat([df1, df2], axis=1)

axis=0 stacks rows (one below the other).

axis=1 stacks columns (side by side).

Examples

Stack two DataFrames with the same columns one below the other. ignore_index=True resets the row numbers.

Pandas

import pandas as pd

df1 = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})
df2 = pd.DataFrame({'A': [5, 6], 'B': [7, 8]})

# Vertical concat (stack rows)
result = pd.concat([df1, df2], axis=0, ignore_index=True)
print(result)

Stack two DataFrames side by side, adding columns.

Pandas

import pandas as pd

df1 = pd.DataFrame({'A': [1, 2]})
df2 = pd.DataFrame({'B': [3, 4]})

# Horizontal concat (stack columns)
result = pd.concat([df1, df2], axis=1)
print(result)

Concatenating with an empty DataFrame just returns the non-empty one.

Pandas

import pandas as pd

# Edge case: empty DataFrame
empty_df = pd.DataFrame()
df = pd.DataFrame({'A': [1, 2]})

result = pd.concat([empty_df, df], axis=0, ignore_index=True)
print(result)

Concatenating DataFrames with only one row each stacks them correctly.

Pandas

import pandas as pd

df_single = pd.DataFrame({'A': [10]})
df_another = pd.DataFrame({'A': [20]})

# Concatenate single-row DataFrames
result = pd.concat([df_single, df_another], axis=0, ignore_index=True)
print(result)

Sample Program

This program creates two small tables for sales in January and February. Then it stacks them vertically to see all sales together.

Pandas

import pandas as pd

# Create two DataFrames with sales data
sales_jan = pd.DataFrame({
    'Product': ['Apple', 'Banana'],
    'Sales': [100, 150]
})

sales_feb = pd.DataFrame({
    'Product': ['Apple', 'Banana'],
    'Sales': [120, 130]
})

print('Sales January:')
print(sales_jan)

print('\nSales February:')
print(sales_feb)

# Combine sales data vertically (stack rows)
all_sales = pd.concat([sales_jan, sales_feb], axis=0, ignore_index=True)

print('\nAll Sales Combined:')
print(all_sales)

Important Notes

Time complexity: Concatenation takes time proportional to the total size of all DataFrames combined.

Space complexity: Requires memory to hold all combined data.

Common mistake: Forgetting ignore_index=True can keep old row numbers, causing duplicate indices.

Use concat() to stack data vertically or horizontally. For merging based on keys, use merge() instead.

Summary

concat() joins DataFrames by stacking rows or columns.

Use axis=0 to stack rows, axis=1 to stack columns.

Remember to reset index with ignore_index=True when stacking rows.