Pandas sort_index() function

Pandas is a Python library, mostly used for data analysis. Pandas make it easier to import, clean, explore, manipulate and analyze data. In this tutorial, we are going to learn about the sort_index() function available in Pandas.

Syntax of the sort_index() function

DataFrame.sort_index(axis=0, level=None, ascending=True, inplace=False, kind='quicksort', na_position='last', sort_remaining=True, ignore_index=False, key=None)

Parameter	Description	Possible Value(s)	Default Value
axis	Sort by the axis along which you want to sort.	0 / ‘index’, 1 / ‘columns’	0
level	Reference the level by which dataframe is to be sorted. If not None, sort on values in specified index level(s).	int or level name or list of ints or list of level names	NA
ascending	Sort in ascending or descending order.	bool or list of bools: ‘True’, ‘False’	True
inplace	Modify the existing dataframe or create a new one.	bool: ‘True’, ‘False’	False
kind	Kind of sorting algorithm to be applied.	‘quiksort’, ‘mergesort’, ‘heapsort’, ‘stable’	quicksort
na_position	The position where to put the NaNs.	‘first’, ‘last’	last
sort_remaining	Specify if the dataframe is to be sorted by other levels after sorting by the specified level.	bool: ‘True’, ‘False’	True
ignore_index	If True, the resulting axis will be labelled 0, 1, …, n – 1.	bool: ‘True’, ‘False’	False
key	Specify if the dataframe is to be sorted on the basis of some custom key	Any custom key defined by the user	NA

Returns: A sorted object i.e. dataframe. If inplace=True, the original dataframe itself is sorted, else the user has to specify a new dataframe for storing the result. If no new dataframe is assigned and inplace=False, then None is returned.

This function works for both numeric as well as non-numeric index values.

Creating a dataframe

Let’s begin with creating a dataframe before we jump to the examples here. This will help us work on a dataframe and understand the implementation of the Python sort_index() function.

import pandas as pd 
import numpy as np

data = {
    'Emp_Name' : ["Anna", "Jeremy", "Dennis", "Courtney", 
                            "Bob", "Ronald"],
    'Dept' : ["Operations", "Technology", "Design", "Administration",
                 "Operations", "Technology"],
    'YOE' : [12, 5, 6, 20, 8, 4]
}

Emp_Id = [4, 3, 6, 7, 1, 2]
emp_data = pd.DataFrame(data, index=Emp_Id)
emp_data

Output:

Sort Index Dataframe — Dataframe for Sort Index Examples

This is a dataframe for storing the information of a company’s employees. It consists of the employee name (Emp_Name), the department in which they work (Dept) and their years of experience (YOE). The data is arranged on the basis of the employee Id (Emp_Id) as specified in the above code.

Examples

Let’s get right into the examples to understand how the sort_index() function is implemented and how it works.

1. Sort in ascending order of the index

# by default, axis=0 so sorting is done by row
emp_data2 = emp_data.sort_index()

emp_data2

Output:

Sort Index Ascending Order — Sort Index in Ascending order

In the above example, we have created a new dataframe named ’emp_data2′ and sorted it basis the index i.e. ‘Emp_Id’ in ascending order. As a result, we got the above dataframe shown in the figure.

2. Sort in descending order of the index

# by default, axis=0 so sorting is done by row
emp_data2.sort_index(ascending=False, inplace=True)

emp_data2

Output:

Sort Index Descending Order — Sort Index in Descending order

Here, we have sorted the dataframe in descending order by specifying ascending=False in the code. As specified, inplace=True and as we have not assigned any new dataframe to the function, emp_data2 is sorted ‘inplace‘.

3. Sort on the basis of column labels in ascending order

# axis=1 implies sort by column
emp_data3 = emp_data.sort_index(axis=1)

emp_data3

Output:

Sort Index By Column in Ascending Order — Sort Index By Column In Ascending Order

In this case, we have a resulting dataframe called ’emp_data3′ in which the column labels are sorted in ascending i.e. lexicographical order. The lexicographic order of the column labels ‘Emp_Name’, ‘Dept’ and ‘YOE’ is ‘Dept’, ‘Emp_Name’, and ‘YOE’, hence the columns are now arranged in this order.

Similarly, we can sort the dataframe in descending order basis the column labels by writing emp_data.sort_index(axis=1, ascending=False). This will result in the below dataframe.

Sort Index By Column in Descending Order — Sort Index By Column In Descending Order

4. Positioning NaNs using the na_position parameter

Let’s consider a slightly modified version of the previous employee dataframe as shown below.

import pandas as pd 
import numpy as np

data = {
    'Emp_Name' : ["Anna", "Jeremy", "Dennis", "Courtney", 
                            "Bob", "Ronald"],
    'Dept' : ["Operations", "Technology", "Design", "Administration", 
                  "Operations", "Technology"],
    'YOE' : [12, 5, 6, 20, 8, 4]
}

Emp_Id = [4, np.nan, 6, np.nan, 1, 2]
emp_df = pd.DataFrame(data, index=Emp_Id)
emp_df

Dataframe With NaN values — Dataframe With NaN

The above dataframe has some NaN values in the index i.e. Emp_Id. Let us see how to position these values.

a. Positioning NaNs first

emp_df2 = emp_df.sort_index(na_position='first')

emp_df2

Output:

Here, the dataframe is sorted on the basis of its index i.e. Emp_Id and the NaN values are placed first in the resulting dataframe.

Also read: How to replace NaN values in a Pandas dataframe with 0?

b. Positioning NaNs last

emp_df3 = emp_df.sort_index(na_position='last')

emp_df3

Output:

In this example, the dataframe is sorted on the basis of its index i.e. Emp_Id and the NaN values are placed last in the resulting dataframe.

Conclusion

That’s all in this tutorial! We have learnt about the sort_index() function in Pandas. You can learn more about Pandas by following our tutorials here.

Reference

sort_index() Official Documentation