NumPy Array Sort – Sorting Arrays Efficiently in Python

Introduction – Why Learn Sorting in NumPy?

Sorting is a core operation in data analysis, used for ranking, organizing, and preprocessing data. Whether you’re preparing inputs for machine learning, ranking scores, or analyzing statistics, NumPy’s fast and flexible sorting tools such as np.sort(), np.argsort(), and sort() method offer optimized ways to work with arrays of any shape.

In this guide, you’ll learn:

• How to sort 1D and multi-dimensional arrays using np.sort()
• The difference between sort(), argsort(), and lexsort()
• How to sort along specific axes
• Performance tips and common pitfalls

Using np.sort() – The Default Sort Function

import numpy as np

arr = np.array([3, 1, 4, 1, 5, 9])
sorted_arr = np.sort(arr)
print(sorted_arr)

Output:

[1 1 3 4 5 9]

np.sort() returns a sorted copy, leaving the original unchanged.

In-Place Sorting with .sort() Method

arr = np.array([3, 1, 4, 1, 5, 9])
arr.sort()
print(arr)

Output:

[1 1 3 4 5 9]

.sort() modifies the original array in-place and is slightly faster.

Sorting 2D Arrays with Axis Control

matrix = np.array([[3, 2, 1], [6, 5, 4]])
sorted_rows = np.sort(matrix, axis=1)  # Sort each row
print(sorted_rows)

Output:

[[1 2 3]
 [4 5 6]]

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

Output:

[[3 2 1]
 [6 5 4]]

axis=1 → sort across columns (row-wise)
axis=0 → sort down rows (column-wise)

Using np.argsort() – Indices that Would Sort the Array

arr = np.array([10, 5, 3])
indices = np.argsort(arr)
print(indices)

Output:

[2 1 0]

Useful when you want to sort related data using these indices:

sorted_arr = arr[indices]

Using np.lexsort() – Sorting by Multiple Keys

names = np.array(['John', 'Bob', 'Alice'])
grades = np.array([90, 80, 80])

# Sort by grade, then name
sorted_idx = np.lexsort((names, grades))
print(names[sorted_idx])

Output:

['Alice' 'Bob' 'John']

lexsort() sorts by the last key first (grade here), then breaks ties with name.

Sorting Boolean and String Arrays

Boolean Arrays

arr = np.array([True, False, True])
print(np.sort(arr))

Output:

[False  True  True]

String Arrays

arr = np.array(['banana', 'apple', 'cherry'])
print(np.sort(arr))

Output:

['apple' 'banana' 'cherry']

Works lexicographically

NumPy Sort Method Comparison

Common Pitfalls

• Confusing sort() with argsort()
• Forgetting that np.sort() doesn’t modify the array
• Using incorrect axis on multi-dimensional arrays
• Assuming argsort() sorts the array – it only gives sort indices

Summary – Key Takeaways

• Use np.sort() for a copy, .sort() for in-place sorting
• Use argsort() to get indices for indirect sorting
• Use lexsort() when sorting by multiple criteria
• Specify axis to control row/column sorting
• Sorting supports numeric, string, and boolean arrays

Real-World Applications

• Sorting prediction probabilities
• Sorting names alphabetically in NLP
• Ranking scores in competitions
• Sorting image pixels for histogram equalization
• Sorting timestamps in chronological order

FAQs – NumPy Array Sort

What’s the difference between sort() and argsort()?
sort() returns sorted data; argsort() returns the index positions.

How do I sort each row of a matrix?
Use:

np.sort(arr, axis=1)

Can I sort in descending order?
Yes:

np.sort(arr)[::-1]  # or use -arr for numeric arrays

How do I sort by two arrays (like keys)?
Use np.lexsort((secondary_key, primary_key))

Can I sort strings and booleans?
Yes. NumPy sorts them in lexicographical or logical order.