The tabular structures in the Python pandas library are also termed DataFrames. 

These structures represent the rows and columns using labels. A row label is called an index, and a column label is called a column index/header.

DataFrames are created by filtering and manipulating large datasets. But these processed DataFrames have the same index as their original datasets. This calls for resetting the index of the DataFrame.

Resetting the index is also helpful:

• In the pre-processing stage, when dropping missing values or filtering data. Besides making the DataFrame smaller, it also jumbles the index.
• When the index labels don't provide much information about the data.
• When the index needs to be treated as a common DataFrame column.

You can use the reset_index() method in pandas to reset the index in a DataFrame. Doing this converts the original index of the DataFrame into a column.

We'll walk you through using the method to reset Pandas DataFrames in this post.

Basics of pandas.reset_index

Besides resetting the index of a DataFrame to a default one, the reset_index() method is also helpful in removing one or more levels of a DataFrame with a MultiIndex.

The syntax of the method is:

pandas.reset_index(level=None, drop=False, inplace=False, col_level=0, col_fill= ”)

Where the parameters have the following meanings:

Parameter	Acceptable Data Types	Default Value	Description
level	int, str, tuple, list		Removes all levels by default. If levels are mentioned, it removes those.
		False	Adds the old index into the DataFrame by default. It does not add it if value is changed to True.
inplace	 bool	False	Carries out the changes in the current DataFrame object
col_level	int, str		Determines which level the labels need to be inserted (if multiple levels are involved). The labels are inserted into the first level (0) by default.
col_fill	object		It determines how the levels are named if columns have multiple levels. If the value is None, the index name is repeated.

The .reset_index() method returns None if inplace = True. Otherwise, the DataFrame with the new index is returned.

Resetting the Index with .reset_index()

Using the .reset_index() method to reset the index is as simple as chaining the method to the DataFrame object. 

You must begin by creating a DataFrame, like so:

import pandas as pd
import numpy as np
import random
# We're making a DataFrame with an initial index. It represents marks out of 50. 

df = pd.DataFrame({
                    'Global Finance': [44, 29, 50, 17, 36],
                    'Politics': [31, 43, 21, 42, 17],
                    'Family Enterprise': [30, 30, 16, 46, 41]
                  }, index=['Leonard', 'Brayan', 'Wendy', 'Nathaniel', 'Edwin']
                 )
df

In tabular form, the DataFrame would look like this:

	Global Finance	Politics	Family Enterprise
Leonard
Brayan
Wendy
Nathaniel
Edwin

Resetting the index is now simply a matter of calling .reset_index(), like so:

df.reset_index()

The index will apply to the DataFrame as a new column named "index." It begins with zero and continues along the length of the DataFrame. It'll look like this:

	index	Global Finance	Politics	Family Enterprise
0	Leonard
1	Brayan
2	Wendy
3	Nathaniel
4	Edwin

Persisting the Change to the DataFrame

The output you see above indicates that the DataFrame's index has been changed. However, if you run "df," you will see that the changes don't persist, and the output does not have an index. 

If you want the changes to persist, you must use the "inplace" parameter, setting its value to "True." Here's what running it looks like:

df.reset_index(inplace=True)

df

The output of this code will be:

	index	Global Finance	Politics	Family Enterprise
0	Leonard
1	Brayan
2	Wendy
3	Nathaniel
4	Edwin

Resetting an Index in a DataFrame with a Named Index

If a DataFrame has a named index, that is, an index with a name, then resetting the index will lead to the named index in question becoming a column name in the DataFrame. 

Let's see how this works by first creating a DataFrame with a named index:

namedIndex = pd.Series(['Leonard', 'Brayan', 'Wendy', 'Nathaniel', 'Edwin'], name='initial_index') # Creating a series and giving it a name

df = pd.DataFrame({
                    'Global Finance': [44, 29, 50, 17, 36],
                    'Politics': [31, 43, 21, 42, 17],
                    'Family Enterprise': [30, 30, 16, 46, 41]
                  }, index=['Leonard', 'Brayan', 'Wendy', 'Nathaniel', 'Edwin']
                 ) # Creating the DataFrame, then passing the named series as the index

df

This DataFrame would look like this:

	Global Finance	Politics	Family Enterprise
initial_index
Leonard
Brayan
Wendy
Nathaniel
Edwin

Executing df.reset_index puts the "initial_index" entry in the table a column name in the DataFrame, like so:

	initial_index	Global Finance	Politics	Family Enterprise
0	Leonard
1	Brayan
2	Wendy
3	Nathaniel
4	Edwin

Resetting a Multi-Level Index in a DataFrame

Let's take a look at a multi-level index in a DataFrame:

# Creating a multi-level index
newIndex = pd.MultiIndex.from_tuples(
                                      [('BBA', 'Leonard'),
                                       ('BBA', 'Brayan'),
                                       ('MBA', 'Wendy'),
                                       ('MBA', 'Nathaniel'),
                                       ('BSC', 'Edwin')
                                      ],
                                  names= ['Branch', 'Name'])

# Creating multi-level columns
columns = pd.MultiIndex.from_tuples(
                                    [('subject1', 'Global Finance'),
                                     ('subject2', 'Politics'),
                                     ('subject3', 'Family Enterprise')
                                    ])

df = pd.DataFrame([
                    (45, 31, 30),
                    (29, 21, 30),
                    (50, 21, 16),
                    (17, 42, 46),
                    (36, 17, 41)      
                    ], index=newIndex, 
                    columns=columns)
df

The output of which is:

		subject1	subject2	subject3
		Global Finance	Politics	Family Enterprise
Branch	Name
BBA	Leonard
Brayan
MBA	Wendy
Nathaniel
BSC	Edwin

The Branch level maps to multiple rows, making this a multi-level index. Applying the .reset_index() function merges the levels as columns in the DataFrame. 

So, running "df.reset_index()" will do this:

			subject1	subject2	subject3
			Global Finance	Politics	Family Enterprise
	Branch	Name
0	BBA	Leonard
1	BBA	Brayan
2	MBA	Wendy
3	MBA	Nathaniel
4	BSC	Edwin

You can also reset the index at the Branch level with the level parameter like so:

df.reset_index(level='Branch')

Which produces the output:

		subject1	subject2	subject3
		Global Finance	Politics	Family Enterprise
Name	Branch
Leonard	BBA
Brayan	BBA
Wendy	MBA
Nathaniel	MBA
Edwin	BSC