Зміст курсу
Optimization Techniques in Python
Optimization Techniques in Python
Using the collections Module
While Python's built-in data types and NumPy arrays will cover most common use cases, the collections module provides specialized data structures with distinct advantages for specific scenarios. The most useful one in terms of performance advantage is deque (double-ended queue).
Unlike lists, which require shifting elements when inserting or removing from the beginning, a deque allows efficient operations on both ends. Therefore, if you frequently need to add or remove elements from either end of a collection, a deque is a better choice.
Here are the methods that perform these operations:
append(): add an element to the right end;appendleft(): add an element to the left end;pop(): remove and return an element from the right end;popleft(): remove and return an element from the left end.
Now, let's compare the performance of list and deque in a practical scenario:
import os os.system('wget https://codefinity-content-media-v2.s3.eu-west-1.amazonaws.com/courses/8d21890f-d960-4129-bc88-096e24211d53/section_1/chapter_3/decorators.py 2>/dev/null') from decorators import timeit_decorator from collections import deque numbers_list = list(range(1, 10000001)) numbers_deque = deque(numbers_list) @timeit_decorator(number=1000) def list_append_left(): # Insert -1 at the beginning numbers_list.insert(0, -1) @timeit_decorator(number=1000) def deque_append_left(): numbers_deque.appendleft(-1) @timeit_decorator(number=1000) def list_pop_left(): # Remove the element at index 0 (first element) numbers_list.pop(0) @timeit_decorator(number=1000) def deque_pop_left(): numbers_deque.popleft() list_append_left() deque_append_left() list_pop_left() deque_pop_left()
In this example, we've created a list and a deque, each containing 1000000 numbers from 1 to 1000000 inclusive. As shown, inserting and removing elements at the beginning is much faster in a deque than in a list and remains efficient regardless of the size.
When it comes to inserting or removing elements at the beginning, both list and deque can handle these tasks efficiently. Therefore, using a deque solely for this purpose often doesn’t provide a significant advantage.
Let's verify it:
import os os.system('wget https://codefinity-content-media-v2.s3.eu-west-1.amazonaws.com/courses/8d21890f-d960-4129-bc88-096e24211d53/section_1/chapter_3/decorators.py 2>/dev/null') from decorators import timeit_decorator from collections import deque numbers_list = list(range(1, 10000001)) numbers_deque = deque(numbers_list) @timeit_decorator(number=1000) def append_right(data_structure): data_structure.append(-1) @timeit_decorator(number=1000) def pop_right(data_structure): data_structure.pop() print('List performance:') append_right(numbers_list) pop_right(numbers_list) print('Dequeue performance:') append_right(numbers_deque) pop_right(numbers_deque)
The performance results are indeed similar for both data structures. However, appending to a list is slightly slower than appending to a deque because lists, implemented as dynamic arrays, occasionally need to resize by allocating a larger memory block and copying elements over. In contrast, deque's block-based structure avoids resizing, making appends consistently faster. This difference, however, becomes noticeable only with relatively large lists.
Дякуємо за ваш відгук!
