Blog

If you’re working with reshaping data in Pandas, you’ve probably come across melt() and pivot(). They’re two powerful but opposite operations — and knowing when to use each is key to structuring your data efficiently.

What is pandas.melt()?

melt() turns a wide-format DataFrame into a long-format DataFrame. It un-pivots your data so that one or more columns become row values.

Example:

import pandas as pd

df = pd.DataFrame({
    'Name': ['Alice', 'Bob'],
    'Math': [85, 90],
    'Science': [88, 95]
})

melted = pd.melt(df, id_vars=['Name'], var_name='Subject', value_name='Score')

Output:

    Name  Subject  Score
0  Alice     Math     85
1    Bob     Math     90
2  Alice  Science     88
3    Bob  Science     95

Use melt() when:

• You want to convert column headers into row values
• You need to prepare data for plotting or exporting
• You’re tidying messy datasets

What is pandas.pivot()?

pivot() does the opposite — it reshapes long-format data into a wide format. You specify which column should be the index, which should be columns, and which should hold values.

Example:

data = {
    'Name': ['Alice', 'Alice', 'Bob', 'Bob'],
    'Subject': ['Math', 'Science', 'Math', 'Science'],
    'Score': [85, 88, 90, 95]
}

df_long = pd.DataFrame(data)
pivoted = df_long.pivot(index='Name', columns='Subject', values='Score')

Output:

Subject  Math  Science
Name                  
Alice      85       88
Bob        90       95

Use pivot() when:

• You have tidy long-format data
• You want to reformat it into a wide summary table
• Your combination of index/columns is unique

Common Pitfalls

• pivot() will raise an error if duplicate index-column pairs exist
• Use pivot_table() instead if you need aggregation or handling duplicates
• melt() is very forgiving and flexible

Summary: melt() vs pivot()

Use melt() to tidy and flatten your data. Use pivot() to rebuild a summary table. These two tools give you full control over reshaping your DataFrames — once you master them, you’ll move between formats with ease.