w3resource

NumPy - Cumulative sum of large array using For loop and optimization


NumPy: Performance Optimization Exercise-11 with Solution


Write a NumPy program that creates a function to calculate the cumulative sum of a large NumPy array using a for loop. Optimize it using NumPy's cumsum() function.

Sample Solution:

Python Code:

import numpy as np

# Generate a large 1D NumPy array with random integers
large_array = np.random.randint(1, 1000, size=1000000)

# Function to calculate cumulative sum using a for loop
def cumulative_sum_with_loop(arr):
    cum_sum = np.empty_like(arr)
    cum_sum[0] = arr[0]
    for i in range(1, len(arr)):
        cum_sum[i] = cum_sum[i - 1] + arr[i]
    return cum_sum

# Calculate cumulative sum using the for loop method
cumsum_with_loop = cumulative_sum_with_loop(large_array)

# Calculate cumulative sum using NumPy's cumsum() function
cumsum_with_numpy = np.cumsum(large_array)

# Display first 10 results to verify
print("First 10 results using for loop:")
print(cumsum_with_loop[:10])

print("First 10 results using NumPy:")
print(cumsum_with_numpy[:10])

Output:

First 10 results using for loop:
[ 654 1458 2169 2433 2841 3397 4373 5060 6027 6726]
First 10 results using NumPy:
[ 654 1458 2169 2433 2841 3397 4373 5060 6027 6726]

Explanation:

  • Importing numpy: We first import the numpy library for array manipulations.
  • Generating a large array: A large 1D NumPy array with random integers is generated.
  • Defining the function: A function cumulative_sum_with_loop is defined to calculate the cumulative sum using a for loop.
  • Calculating with loop: The cumulative sum is calculated using the for loop method.
  • Calculating with numpy: The cumulative sum is calculated using NumPy's built-in cumsum() function.
  • Displaying results: The first 10 results from both methods are printed out to verify correctness.

Python-Numpy Code Editor: