How to Flatten Multi-Index in pandas: Simple Guide

To flatten a pandas multi-index DataFrame, use reset_index() to convert index levels into columns. Alternatively, combine index levels into single strings with map or join on the index names.

Syntax

The main method to flatten a multi-index in pandas is reset_index(). It moves the index levels into regular columns, making the DataFrame easier to work with.

Another way is to combine the multi-index levels into one by joining their names with a separator using map and join.

python

df.reset_index(level=None, drop=False, inplace=False, col_level=0, col_fill='')

Example

This example shows how to flatten a multi-index DataFrame by resetting the index and by joining index levels into a single level.

python

import pandas as pd

# Create a sample multi-index DataFrame
data = {"value": [10, 20, 30, 40]}
index = pd.MultiIndex.from_tuples(
    [("A", 1), ("A", 2), ("B", 1), ("B", 2)],
    names=["letter", "number"]
)
df = pd.DataFrame(data, index=index)

# Flatten by resetting index
flat_df = df.reset_index()

# Flatten by joining index levels into one string
flat_index = df.index.map(lambda x: f"{x[0]}_{x[1]}")
df_flat_index = df.copy()
df_flat_index.index = flat_index

flat_df, df_flat_index

Output

   letter  number  value
0      A       1     10
1      A       2     20
2      B       1     30
3      B       2     40

        value
A_1      10
A_2      20
B_1      30
B_2      40

Common Pitfalls

One common mistake is forgetting that reset_index() returns a new DataFrame by default and does not change the original unless inplace=True is set.

Another pitfall is losing the original index if drop=True is used unintentionally.

python

import pandas as pd

# Sample multi-index DataFrame
data = {"value": [1, 2]}
index = pd.MultiIndex.from_tuples([("X", "a"), ("Y", "b")], names=["first", "second"])
df = pd.DataFrame(data, index=index)

# Wrong: reset_index without assignment or inplace
_ = df.reset_index()
print(df)  # Original df unchanged

# Right: assign back or use inplace
flat_df = df.reset_index()
print(flat_df)

# Or inplace
# df.reset_index(inplace=True)
# print(df)

Output

              value
first second       
X     a           1
Y     b           2

  first second  value
0     X      a      1
1     Y      b      2

Quick Reference

Method	Description	Example
reset_index()	Moves multi-index levels to columns	df.reset_index()
index.map()	Combine multi-index levels into one string	df.index.map(lambda x: f"{x[0]}_{x[1]}")
reset_index(drop=True)	Remove index and do not keep as columns	df.reset_index(drop=True)

Key Takeaways

Use df.reset_index() to flatten multi-index by turning index levels into columns.

Remember reset_index returns a new DataFrame unless inplace=True is set.

You can join multi-index levels into one string to create a flat index.

Avoid losing data by not using drop=True unless you want to remove index levels.

Flattening multi-index makes data easier to analyze and visualize.