Mastering NumPy in Python: The Backbone of Scientific Computing

Chapter 4: Advanced Indexing, Masking, and Manipulation in NumPy

🔹 1. Introduction

Indexing and slicing are fundamental when working with arrays in NumPy. While basic indexing is similar to Python lists, NumPy extends it with powerful features like boolean masking, fancy indexing, and efficient reshaping. These techniques allow you to extract, modify, and analyze subsets of data more intuitively and performantly.

In this chapter, we’ll explore:

• Boolean indexing
• Fancy (advanced) indexing
• Slicing tips
• Combining and splitting arrays
• Copy vs. view behavior

Mastering these techniques is essential for efficient data manipulation, preprocessing, and modeling tasks in data science, machine learning, and beyond.

🔹 2. Basic Indexing Review

Before diving deep, let’s recall basic indexing:

import numpy as np

arr = np.array([[10, 20, 30],

                [40, 50, 60]])

print(arr[1, 2])  # Output: 60

• arr[i, j]: Row i, column j
• Negative indexing works: arr[-1, -1] → 60

🔹 3. Slicing Arrays

✅ 1D Slicing

arr = np.array([0, 1, 2, 3, 4, 5])

print(arr[1:4])     # [1 2 3]

print(arr[::-1])    # Reversed array

✅ 2D Slicing

arr = np.array([[10, 20, 30], [40, 50, 60], [70, 80, 90]])

print(arr[:2, 1:])   # [[20 30] [50 60]]

🔹 4. Boolean Indexing (Masking)

Allows selection based on condition:

arr = np.array([5, 10, 15, 20])

mask = arr > 10

print(arr[mask])  # [15 20]

Or inline:

print(arr[arr < 15])  # [5 10]

✅ Use Case: Filter Even Numbers

arr = np.arange(10)

even = arr[arr % 2 == 0]  # [0 2 4 6 8]

🔹 5. Fancy Indexing

Fancy indexing allows you to access multiple indices at once using lists or arrays.

arr = np.array([100, 200, 300, 400, 500])

print(arr[[0, 2, 4]])  # [100 300 500]

You can also use it with 2D arrays:

arr2d = np.array([[1, 2], [3, 4], [5, 6]])

print(arr2d[[0, 2], [1, 0]])  # [2 5]

🔹 6. Combining Boolean & Fancy Indexing

You can chain conditions and use logical operators:

arr = np.array([5, 10, 15, 20, 25])

print(arr[(arr > 10) & (arr < 25)])  # [15 20]

Note: Use &, |, ~ instead of and, or, not for arrays.

🔹 7. Indexing with np.where()

np.where() returns indices or replaces values based on a condition.

✅ Get Indices

arr = np.array([1, 3, 7, 9])

indices = np.where(arr > 5)

print(indices)  # (array([2, 3]),)

✅ Conditional Assignment

arr = np.array([10, 15, 20])

new_arr = np.where(arr > 12, 1, 0)

print(new_arr)  # [0 1 1]

🔹 8. Copy vs. View: What’s the Difference?

✅ View (shallow copy)

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

b = a[1:]

b[0] = 100

print(a)  # [1 100 3] → View affects original

✅ Copy (deep copy)

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

b = a[1:].copy()

b[0] = 100

print(a)  # [1 2 3] → Original unchanged

🔹 9. Array Manipulation: Stack, Split, Reshape

✅ Reshape

arr = np.arange(6)

arr2d = arr.reshape(2, 3)

✅ Stacking Arrays

a = np.array([1, 2])

b = np.array([3, 4])

np.vstack((a, b))  # Vertical stack

np.hstack((a, b))  # Horizontal stack

✅ Splitting Arrays

arr = np.array([[1, 2, 3], [4, 5, 6]])

np.hsplit(arr, 3)  # Split into 3 columns

np.vsplit(arr, 2)  # Split into 2 rows

🔹 10. Summary Table