How to Flatten MultiIndex in Pandas (With Examples)


You can use the following basic syntax to flatten a MultiIndex in pandas:

#flatten all levels of MultiIndex
df.reset_index(inplace=True)

#flatten specific levels of MultiIndex
df.reset_index(inplace=True, level = ['level_name'])

The following examples show how to use this syntax in practice.

Example 1: Flatten All Levels of MultiIndex in Pandas

Suppose we have the following MultiIndex pandas DataFrame:

import pandas as pd

#create DataFrame
index_names = pd.MultiIndex.from_tuples([('Level1','Lev1', 'L1'),
                                       ('Level2','Lev2', 'L2'),
                                       ('Level3','Lev3', 'L3'),
                                       ('Level4','Lev4', 'L4')],
                                       names=['Full','Partial', 'ID'])

data = {'Store': ['A','B','C','D'],
        'Sales': [12, 44, 29, 35]}

df = pd.DataFrame(data, columns = ['Store','Sales'], index=index_names)

#view DataFrame
df

			Store	Sales
Full	Partial	ID		
Level1	Lev1	L1	A	17
Level2	Lev2	L2	B	22
Level3	Lev3	L3	C	29
Level4	Lev4	L4	D	35

We can use the following syntax to flatten every level of the MultiIndex into columns in the DataFrame:

#flatten every level of MultiIndex 
df.reset_index(inplace=True)

#view updated DataFrame
df

        Full	Partial	ID	Store	Sales
0	Level1	Lev1	L1	A	12
1	Level2	Lev2	L2	B	44
2	Level3	Lev3	L3	C	29
3	Level4	Lev4	L4	D	35

Notice that each level of the MultiIndex is now a column in the DataFrame.

Example 2: Flatten Specific Levels of MultiIndex in Pandas

Suppose we have the same pandas DataFrame as the previous example:

#view DataFrame
df

			Store	Sales
Full	Partial	ID		
Level1	Lev1	L1	A	12
Level2	Lev2	L2	B	44
Level3	Lev3	L3	C	29
Level4	Lev4	L4	D	35

The following code shows how to flatten just one specific level of the MultiIndex:

#flatten 'ID' level only
df.reset_index(inplace=True, level = ['ID'])

#view updated DataFrame
df

		ID	Store	Sales
Full	Partial			
Level1	Lev1	L1	A	12
Level2	Lev2	L2	B	44
Level3	Lev3	L3	C	29
Level4	Lev4	L4	D	35

And the following code shows how to flatten several specific levels of the MultiIndex:

#flatten 'ID' level only
df.reset_index(inplace=True, level = ['Partial', 'ID'])

#view updated DataFrame
df

	Partial	 ID	Store	Sales
Full				
Level1	Lev1	 L1	A	12
Level2	Lev2	 L2	B	44
Level3	Lev3	 L3	C	29
Level4	Lev4	 L4	D	35

Additional Resources

The following tutorials explain how to perform other common functions in pandas:

How to Convert Index to Column in Pandas
How to Rename Index in Pandas
How to Set Column as Index in Pandas

Leave a Reply

Your email address will not be published.