Why indexing matters in NumPy - Performance Analysis

Time Complexity: Why indexing matters

O(k)

Understanding Time Complexity

We want to see how fast numpy operations run when we pick parts of data using indexing.

How does the way we select data affect the time it takes?

Scenario Under Consideration

Analyze the time complexity of the following code snippet.

import numpy as np

arr = np.arange(1000000)

# Access a single element
single = arr[500000]

# Access a slice
slice_ = arr[100000:200000]

# Access with a boolean mask
mask = arr % 2 == 0
filtered = arr[mask]

This code shows three ways to get data from a large array: one element, a slice, and a filtered subset.

Identify Repeating Operations

Look at what repeats when we pick data.

• Primary operation: Accessing elements by index or condition.
• How many times: For single element, once; for slice, many times equal to slice size; for mask, once per element to check condition.

How Execution Grows With Input

How does time grow when array size grows?

Input Size (n)	Approx. Operations
10	Single: 1, Slice: up to 10, Mask: 10
100	Single: 1, Slice: up to 100, Mask: 100
1000	Single: 1, Slice: up to 1000, Mask: 1000

Picking one element stays fast no matter size. Picking slices or filtered data grows with how much we take or check.

Final Time Complexity

Time Complexity: O(k) where k is the number of elements accessed or checked

This means time grows with how many elements you pick or test, not the total array size if you pick just one.

Common Mistake

[X] Wrong: "Accessing any element in numpy always takes the same time regardless of how many elements we get."

[OK] Correct: Accessing one element is fast, but getting many elements or filtering means more work and more time.

Interview Connect

Knowing how indexing affects speed helps you write faster code and explain your choices clearly in real projects.

Self-Check

"What if we used fancy indexing with a list of random indices? How would the time complexity change?"