Sorting Arrays & Extracting Indices (np.sort vs np.argsort)0%

Searching & Counting (argmax, argmin, nonzero, count_nonzero)

Sorting Arrays & Extracting Indices (np.sort vs np.argsort)

Beginner12 min read•Updated: Jul 11, 2026

Sorting Arrays: np.sort() vs np.argsort()

Sorting data is a core algorithmic primitive in numerical computing. In scientific pipelines, sorting enables ranking, percentile calculation, quantile binning, nearest-neighbor searches, and ordering multi-column tables.

NumPy provides two complementary sorting paradigms:

Direct Sorting (np.sort() / arr.sort()): Reorders and returns the actual values.
Indirect Sorting (np.argsort()): Returns the indices that would sort the array, enabling simultaneous sorting of multiple associated datasets.

1. Direct Sorting: Function vs In-Place Method

np.sort(arr): Returns a sorted copy of the array, leaving the original array completely untouched.
arr.sort(): Sorts the array in-place directly within its existing memory buffer, returning None.

1import numpy as np
2 
3numbers = np.array([42, 17, 89, 3, 25, 68])
4 
5# 1. np.sort() returns a new sorted array:
6sorted_copy = np.sort(numbers)
7print("Sorted copy:    ", sorted_copy)
8print("Original intact:", numbers)
9 
10# 2. arr.sort() mutates in-place:
11numbers.sort()
12print("After in-place: ", numbers)

2. Multi-Dimensional Sorting Along Axes

When sorting multi-dimensional arrays, you can specify the target axis along which elements are ordered:

1matrix = np.array([
2    [9, 2, 7],
3    [5, 8, 1],
4    [4, 6, 3]
5])
6 
7# Sort across columns (within each row) -> axis=1 (or axis=-1, default)
8row_sorted = np.sort(matrix, axis=1)
9print("Row-wise sorted (axis=1):
10", row_sorted)
11 
12# Sort down rows (within each column) -> axis=0
13col_sorted = np.sort(matrix, axis=0)
14print("
15Column-wise sorted (axis=0):
16", col_sorted)

Output:

1Row-wise sorted (axis=1):
2 [[2 7 9]
3 [1 5 8]
4 [3 4 6]]
5 
6Column-wise sorted (axis=0):
7 [[4 2 1]
8 [5 6 3]
9 [9 8 7]]

3. Indirect Sorting with `np.argsort()`

In real-world data science, you rarely sort a single isolated column. Instead, you have parallel arrays—such as employee names, salaries, and performance ratings—and sorting by salary must reorder the names and ratings in identical order.

np.argsort() returns the array of integer indices that put the array into sorted order:

1names = np.array(["Alice", "Bob", "Charlie", "Diana", "Evan"])
2salaries = np.array([85000, 42000, 110000, 67000, 95000])
3 
4# Get sorted indices of salaries (ascending)
5sorted_order = np.argsort(salaries)
6print("Indices that sort salaries:", sorted_order)
7# Output: [1, 3, 0, 4, 2] -> index 1 (Bob: 42k) is smallest, index 2 (Charlie: 110k) is largest
8 
9# Reorder both arrays simultaneously using fancy indexing:
10print("
11Employees ranked by salary (Lowest to Highest):")
12for idx in sorted_order:
13    print(f"  {names[idx]:10s} : ${salaries[idx]:,}")

Sorting in Descending Order:

To sort in descending order (highest first), slice the resulting indices with [::-1]:

1descending_order = np.argsort(salaries)[::-1]
2print("Top earner:", names[descending_order[0]])  # Charlie

4. Multi-Key Lexicographical Sorting with `np.lexsort()`

When sorting by multiple columns (e.g. sort primarily by Surname, and secondarily by First Name):

Note

Important Syntax Note for np.lexsort(): The keys are passed as a tuple, and the LAST key in the tuple is the PRIMARY sort key!

1first_names = np.array(["John", "Jane", "Alice", "Bob"])
2last_names  = np.array(["Smith", "Smith", "Brown", "Brown"])
3 
4# Sort by last name (primary), then first name (secondary)
5# Pass (secondary, primary)
6sort_idx = np.lexsort((first_names, last_names))
7 
8print("Sorted by Last Name, then First Name:")
9for idx in sort_idx:
10    print(f"  {last_names[idx]}, {first_names[idx]}")

Output:

1Sorted by Last Name, then First Name:
Brown, Alice
Brown, Bob
Smith, Jane
Smith, John

Interactive Quiz• Knowledge Check

Multiple Choice Questions

Question 1

What is the return value of arr.sort()?

Question 2

What does np.argsort(arr) return?

Question 3

Given salaries = np.array([50000, 30000, 80000]), what is the output of np.argsort(salaries)?

Question 4

In np.lexsort((col_b, col_a)), which column acts as the primary sort key?

Question 5

How can you sort a 1D array in descending order using np.sort()?

Next Lesson

Searching & Counting (argmax, argmin, nonzero, count_nonzero)

Continue learning with hands-on practice, examples, and exercises in the upcoming topic.

Related Lessons

Previous Lesson	Next Lesson
Directional Aggregations with the axis Parameter	Searching & Counting (argmax, argmin, nonzero, count_nonzero)

Practice Quiz

Test your understanding of this lesson with 5 questions. Each question has one correct answer.

Previous Next

PreviousPrevNext

MSK Institute • Shikohabad

Loading Page Content...

Preparing courses, live labs, and learning resources.सामग्री लोड हो रही है, कृपया प्रतीक्षा करें...

Connecting serverPlease wait...

💻 Practical Labs🚀 Free Demo📜 Verified Certificates

Searching & Counting (argmax, argmin, nonzero, count_nonzero)

Sorting Arrays & Extracting Indices (np.sort vs np.argsort)

Beginner12 min read•Updated: Jul 11, 2026

Study Materials Cheatsheet

Sorting Arrays: np.sort() vs np.argsort()

NumPy provides two complementary sorting paradigms:

Direct Sorting (np.sort() / arr.sort()): Reorders and returns the actual values.
Indirect Sorting (np.argsort()): Returns the indices that would sort the array, enabling simultaneous sorting of multiple associated datasets.

1. Direct Sorting: Function vs In-Place Method

np.sort(arr): Returns a sorted copy of the array, leaving the original array completely untouched.
arr.sort(): Sorts the array in-place directly within its existing memory buffer, returning None.

1import numpy as np
2 
3numbers = np.array([42, 17, 89, 3, 25, 68])
4 
5# 1. np.sort() returns a new sorted array:
6sorted_copy = np.sort(numbers)
7print("Sorted copy:    ", sorted_copy)
8print("Original intact:", numbers)
9 
10# 2. arr.sort() mutates in-place:
11numbers.sort()
12print("After in-place: ", numbers)

2. Multi-Dimensional Sorting Along Axes

When sorting multi-dimensional arrays, you can specify the target axis along which elements are ordered:

1matrix = np.array([
2    [9, 2, 7],
3    [5, 8, 1],
4    [4, 6, 3]
5])
6 
7# Sort across columns (within each row) -> axis=1 (or axis=-1, default)
8row_sorted = np.sort(matrix, axis=1)
9print("Row-wise sorted (axis=1):
10", row_sorted)
11 
12# Sort down rows (within each column) -> axis=0
13col_sorted = np.sort(matrix, axis=0)
14print("
15Column-wise sorted (axis=0):
16", col_sorted)

Output:

1Row-wise sorted (axis=1):
2 [[2 7 9]
3 [1 5 8]
4 [3 4 6]]
5 
6Column-wise sorted (axis=0):
7 [[4 2 1]
8 [5 6 3]
9 [9 8 7]]

3. Indirect Sorting with `np.argsort()`

np.argsort() returns the array of integer indices that put the array into sorted order:

1names = np.array(["Alice", "Bob", "Charlie", "Diana", "Evan"])
2salaries = np.array([85000, 42000, 110000, 67000, 95000])
3 
4# Get sorted indices of salaries (ascending)
5sorted_order = np.argsort(salaries)
6print("Indices that sort salaries:", sorted_order)
7# Output: [1, 3, 0, 4, 2] -> index 1 (Bob: 42k) is smallest, index 2 (Charlie: 110k) is largest
8 
9# Reorder both arrays simultaneously using fancy indexing:
10print("
11Employees ranked by salary (Lowest to Highest):")
12for idx in sorted_order:
13    print(f"  {names[idx]:10s} : ${salaries[idx]:,}")

Sorting in Descending Order:

To sort in descending order (highest first), slice the resulting indices with [::-1]:

1descending_order = np.argsort(salaries)[::-1]
2print("Top earner:", names[descending_order[0]])  # Charlie

4. Multi-Key Lexicographical Sorting with `np.lexsort()`

When sorting by multiple columns (e.g. sort primarily by Surname, and secondarily by First Name):

Note

Important Syntax Note for np.lexsort(): The keys are passed as a tuple, and the LAST key in the tuple is the PRIMARY sort key!

1first_names = np.array(["John", "Jane", "Alice", "Bob"])
2last_names  = np.array(["Smith", "Smith", "Brown", "Brown"])
3 
4# Sort by last name (primary), then first name (secondary)
5# Pass (secondary, primary)
6sort_idx = np.lexsort((first_names, last_names))
7 
8print("Sorted by Last Name, then First Name:")
9for idx in sort_idx:
10    print(f"  {last_names[idx]}, {first_names[idx]}")

Output:

1Sorted by Last Name, then First Name:
Brown, Alice
Brown, Bob
Smith, Jane
Smith, John

Interactive Quiz• Knowledge Check

Multiple Choice Questions

Question 1

What is the return value of arr.sort()?

Question 2

What does np.argsort(arr) return?

Question 3

Given salaries = np.array([50000, 30000, 80000]), what is the output of np.argsort(salaries)?

Question 4

In np.lexsort((col_b, col_a)), which column acts as the primary sort key?

Question 5

How can you sort a 1D array in descending order using np.sort()?

Next Lesson

Searching & Counting (argmax, argmin, nonzero, count_nonzero)

Continue learning with hands-on practice, examples, and exercises in the upcoming topic.

Related Lessons

Previous Lesson	Next Lesson
Directional Aggregations with the axis Parameter	Searching & Counting (argmax, argmin, nonzero, count_nonzero)

Practice Quiz

Test your understanding of this lesson with 5 questions. Each question has one correct answer.

Previous Next

Sorting Arrays: np.sort() vs np.argsort()

1. Direct Sorting: Function vs In-Place Method

2. Multi-Dimensional Sorting Along Axes

3. Indirect Sorting with np.argsort()

Sorting in Descending Order:

4. Multi-Key Lexicographical Sorting with np.lexsort()

Multiple Choice Questions

Searching & Counting (argmax, argmin, nonzero, count_nonzero)

Related Lessons

Practice Quiz

Loading Page Content...

Sorting Arrays: np.sort() vs np.argsort()

1. Direct Sorting: Function vs In-Place Method

2. Multi-Dimensional Sorting Along Axes

3. Indirect Sorting with np.argsort()

Sorting in Descending Order:

4. Multi-Key Lexicographical Sorting with np.lexsort()

Multiple Choice Questions

Searching & Counting (argmax, argmin, nonzero, count_nonzero)

Related Lessons

Practice Quiz

3. Indirect Sorting with `np.argsort()`

4. Multi-Key Lexicographical Sorting with `np.lexsort()`

3. Indirect Sorting with `np.argsort()`

4. Multi-Key Lexicographical Sorting with `np.lexsort()`