a.sort, sorted(a), np_argsort(a) and np.lexsort(b, a) in Python

Python provides built-in sorting functions (sorted(), list.sort()) and NumPy provides advanced sorting (np.argsort(), np.lexsort()) for working with arrays and multiple sort keys.

sorted()

Returns a new sorted list without modifying the original ?

a = [9, 5, 3, 1, 12, 6]
b = sorted(a)
print("Sorted Array:", b)
print("Original Array:", a)

Sorted Array: [1, 3, 5, 6, 9, 12]
Original Array: [9, 5, 3, 1, 12, 6]

list.sort()

Sorts the list in-place (modifies the original, returns None). Faster than sorted() since it doesn't create a copy ?

a = [9, 5, 3, 1, 12, 6]
print("Original Array:", a)
a.sort()
print("Sorted Array:", a)

Original Array: [9, 5, 3, 1, 12, 6]
Sorted Array: [1, 3, 5, 6, 9, 12]

numpy.argsort()

Returns the indices that would sort the array, not the sorted values themselves. Useful when you need to track original positions ?

import numpy as np

x = np.array([9, 5, 3, 1, 12, 6])
print("Array:", x)

# Print index-value pairs
for i in range(len(x)):
    print(f"  [{i}] = {x[i]}")

# Get indices of sorted elements
s = np.argsort(x)
print("Sorted indices:", s)
print("Sorted values:", x[s])

Array: [ 9  5  3  1 12  6]
  [0] = 9
  [1] = 5
  [2] = 3
  [3] = 1
  [4] = 12
  [5] = 6
Sorted indices: [3 2 1 5 0 4]
Sorted values: [ 1  3  5  6  9 12]

Index 3 (value 1) comes first, then index 2 (value 3), and so on.

numpy.lexsort()

Sorts using multiple keys (like sorting a spreadsheet by column A, then by column B for ties). The last key in the argument is the primary sort key ?

import numpy as np

colA = [2, 5, 1, 8, 1]   # Primary sort key
colB = [9, 0, 3, 2, 0]   # Secondary sort key

# Sort by colA first, then colB for ties
sorted_index = np.lexsort((colB, colA))
print("Sorted indices:", sorted_index)
print("Sorted pairs:", [(colA[i], colB[i]) for i in sorted_index])

Sorted indices: [4 2 0 1 3]
Sorted pairs: [(1, 0), (1, 3), (2, 9), (5, 0), (8, 2)]

ColA has two values of 1 at indices 2 and 4. Since colB[4]=0 < colB[2]=3, index 4 comes before index 2 in the result.

Comparison

Conclusion

Use sorted() when you need a new sorted list and list.sort() for in-place sorting. For NumPy arrays, np.argsort() returns sorted indices to track positions, and np.lexsort() handles multi-key sorting where the last argument is the primary key.