How to Sort Array in NumPy: Simple Syntax and Examples

To sort an array in NumPy, use np.sort(array) to return a sorted copy or array.sort() to sort the array in place. Both methods sort elements in ascending order by default.

Syntax

NumPy provides two main ways to sort arrays:

• np.sort(array, axis=-1, kind='quicksort', order=None): Returns a sorted copy of the array.
• array.sort(axis=-1, kind='quicksort', order=None): Sorts the array in place, modifying the original array.

Parameters:

• array: The NumPy array to sort.
• axis: The axis along which to sort. Default is the last axis.
• kind: Sorting algorithm ('quicksort', 'mergesort', 'heapsort', 'stable').
• order: When sorting structured arrays, specifies field names.

python

import numpy as np

# Using np.sort returns a sorted copy
sorted_array = np.sort(np.array([3, 1, 2]))

# Using ndarray.sort sorts in place
arr = np.array([3, 1, 2])
arr.sort()

Example

This example shows how to sort a 1D and 2D NumPy array using both np.sort and array.sort(). It also demonstrates sorting along different axes.

python

import numpy as np

# 1D array
arr1d = np.array([5, 2, 9, 1])
sorted_copy = np.sort(arr1d)

# Original array remains unchanged

# Sort in place
arr1d.sort()

# 2D array
arr2d = np.array([[3, 7, 5], [8, 4, 6]])

# Sort each row (axis=1)
sorted_rows = np.sort(arr2d, axis=1)

# Sort each column (axis=0)
sorted_cols = np.sort(arr2d, axis=0)

sorted_copy, arr1d, sorted_rows, sorted_cols

Output

(
 array([1, 2, 5, 9]),
 array([1, 2, 5, 9]),
 array([[3, 5, 7],
        [4, 6, 8]]),
 array([[3, 4, 5],
        [8, 7, 6]])
)

Common Pitfalls

1. Confusing np.sort() and array.sort(): np.sort() returns a new sorted array, leaving the original unchanged, while array.sort() changes the original array.

2. Forgetting to assign the result of np.sort(): Since it returns a new array, you must save it to a variable.

3. Sorting multi-dimensional arrays without specifying axis: By default, sorting happens along the last axis, which may not be what you expect.

python

import numpy as np

arr = np.array([3, 1, 2])

# Wrong: does not change arr
np.sort(arr)
print(arr)  # Output: [3 1 2]

# Right: assign sorted array
arr_sorted = np.sort(arr)
print(arr_sorted)  # Output: [1 2 3]

# Sort in place
arr.sort()
print(arr)  # Output: [1 2 3]

Output

[3 1 2]
[1 2 3]
[1 2 3]

Quick Reference

Here is a quick summary of sorting arrays in NumPy:

Function/Method	Description	Modifies Original Array?
np.sort(array, axis=-1)	Returns a sorted copy of the array	No
array.sort(axis=-1)	Sorts the array in place	Yes

Key Takeaways

Use np.sort(array) to get a sorted copy without changing the original array.

Use array.sort() to sort the array in place and modify the original data.

Specify the axis parameter to control sorting direction in multi-dimensional arrays.

Remember to assign the result of np.sort() if you want to keep the sorted array.

Choose the sorting algorithm with the kind parameter if needed, default is quicksort.