NumPy np.argsort()

The argsort() function in NumPy allows you to perform an indirect sorting of an array along a specified axis using the set algorithm.

Let us explore how to work with the NumPy argsort() function.

NumPy argsort() Function Syntax

The function syntax and parameters are as shown in the following:

numpy.argsort(a, axis=- 1, kind=None, order=None)

The parameters are defined according to their following functions:

  1. a – refers to the input array.
  2. axis – specifies along which axis to sort the input array. The value is set to -1, which uses the last index. Set the axis parameter to None if you want to flatten the array.
  3. kind – specifies the sorting algorithm. Acceptable values include ‘quicksort’, ‘mergesort’, ‘heapsort’, and ‘stable’. By default, the function will use a quicksort sorting algorithm.
  4. order – defines the order under which to compare the fields.

Function Return Value
The function returns an array of indices of the sorted array according to the set parameters.

Example 1 – Flattened Array Sort

The following example shows how to flatten and sort the array by specifying the axis parameter as “None”.

# import numpy
import numpy as np
arr = [0,1,5,3,9,6,7,4,2,8]
print("unsorted array: ", arr)
print(f"sorted array: {np.argsort(arr, axis=None)}")

The previous code returns an array with the indices of the sorted array. An example output is as shown below:

unsorted array:  [0, 1, 5, 3, 9, 6, 7, 4, 2, 8]
sorted array: [0 1 8 3 7 2 5 6 9 4]

Example 2 – Flattened Array Sort (Different Algorithms)

Consider the example below that shows how to sort a flattened array using various sorting algorithms:

arr = [0,1,5,3,9,6,7,4,2,8]
quick_sort = np.argsort(arr, axis=None, kind='quicksort')
merge_sort = np.argsort(arr, axis=None, kind='mergesort')
heap_sort = np.argsort(arr, axis=None, kind='heapsort')
stable = np.argsort(arr, axis=None, kind='heapsort')

print(f" quicksort: {quick_sort}\n mergesort: {merge_sort}\n heapsort: {heap_sort}\n stable: {stable}")

The previous code utilizes various sorting algorithms and returns the resulting array indices.

The following is an output example:

 quicksort: [0 1 8 3 7 2 5 6 9 4]
 mergesort: [0 1 8 3 7 2 5 6 9 4]
 heapsort: [0 1 8 3 7 2 5 6 9 4]
 stable: [0 1 8 3 7 2 5 6 9 4]

Example 3 – Sort 2D Array Along Various Axis

Let’s look at an example in sorting a 2D array along the various axis.

print(f"axis 0: {np.argsort(arr, axis=0)}")

The return array is as follows:

axis 0: [[1 1 1]
 [0 0 0]]

To sort along axis 1, apply the following code:

print(f"axis 1: {np.argsort(arr, axis=1)}")

This should return to:

axis 1: [[2 1 0]
 [2 1 0]]


This article serves as a guide on how to use the argsort() function in NumPy to sort an array along a specific axis using various algorithms. More tutorials are available at Linux Hint.

About the author

John Otieno

My name is John and am a fellow geek like you. I am passionate about all things computers from Hardware, Operating systems to Programming. My dream is to share my knowledge with the world and help out fellow geeks. Follow my content by subscribing to LinuxHint mailing list