NumPy - argsort() function

The NumPy argsort() function returns the indices that would sort an array.

Syntax

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

Parameters

`a`	`Required.` Specify the array (array_like) to be sorted.
`axis`	`Optional.` Specify the axis along which to sort. If None, the array is flattened before sorting. Default is -1, which sorts along the last axis.
`kind`	`Optional.` Specify sorting algorithm. It can take values from {'quicksort', 'mergesort', 'heapsort', 'stable'}. Default: 'quicksort'
`order`	`Optional.` Specify string or list of strings containing fields. When a is an array with fields defined, this argument specifies the order in which the fields need to the compared.

Return Value

Returns an array of indices that sort a along the specified axis. If a is one-dimensional, a[index_array] yields a sorted a. In more general, np.take_along_axis(a, index_array, axis=axis) always yields the sorted a, irrespective of dimensionality.

Example:

In the example below, argsort() function is used to get the indices which is further used to yield the sorted array.

import numpy as np
Arr = np.array([50, 40, 10, 60, 30, 20])

#displaying the array
print("The original array:")
print(Arr)

#getting the indices from argsort()
x = np.argsort(Arr)
print("\nIndices from argsort():")
print(x)

#yielding sorted array
print("\nSorted array:")
print(Arr[x])

The output of the above code will be:

The original array:
[50 40 10 60 30 20]

Indices from argsort():
[2 5 4 1 0 3]

Sorted array:
[10 20 30 40 50 60]

Example: using axis parameter

The axis parameter can be used to get the indices that will sort the array along given axis, as shown in the example below:

import numpy as np
Arr = np.array([[40,10],
                [20,50]])

#displaying the array
print("The original array:")
print(Arr)

#yielding sorted array along axis=0
x = np.argsort(Arr, axis=0)
print("\nSorted array along axis=0:")
print(np.take_along_axis(Arr, x, axis=0))

#yielding sorted array along axis=1
y = np.argsort(Arr, axis=1)
print("\nSorted array along axis=1:")
print(np.take_along_axis(Arr, y, axis=1))

The output of the above code will be:

The original array:
[[40 10]
 [20 50]]

Sorted array along axis=0:
[[20 10]
 [40 50]]

Sorted array along axis=1:
[[10 40]
 [20 50]]

Example: using order parameter

The order parameter can be used to specify which field need to sort first. Consider the example below:

import numpy as np
Arr = np.array([(20,60),
                (20,50),
                (20,55),
                (10,75),
                (10,25),
                (10,50)], 
                dtype=[('x', '<i4'), ('y', '<i4')])

#displaying the array
print("The original array:")
print(Arr)

#yielding sorted array
indices = np.argsort(Arr, order=('x','y') )
print("\nSorted array:")
print(Arr[indices])

The output of the above code will be:

The original array:
[(20, 60) (20, 50) (20, 55) (10, 75) (10, 25) (10, 50)]

Sorted array:
[(10, 25) (10, 50) (10, 75) (20, 50) (20, 55) (20, 60)]

❮ NumPy - Functions