4.6 KiB
4.6 KiB
List vs tuple
Points communs
list et tuple sont des données de type sequence :
- L'ordre d'insertion des données est conservé (ordered sets),
- Support des fonctions de concaténation, répétition, indexation et slicing,
- Stockage de données hétérogèes (généralement homogènes)/
Différences
- Syntaxe ([] pour list, () pour tuple),
- La list est mutable, le tuple immutable (ne peut être modifié une fois créé),
Empreinte RAM
- Le tuple requiert moins de RAM :
from sys import getsizeof
from typing import List, Tuple
def print_list_tuple_comparison(list_: List, tuple_: Tuple) -> None:
list_size = getsizeof(list_)
tuple_size = getsizeof(tuple_)
diff = list_size - tuple_size
print(f"Number of elements: {len(list_)}, list size: {list_size:<3}, tuple size: {tuple_size:<3}, diff: {diff} bytes ({diff/list_size*100:.2f}%)")
def print_list_tuple_ram_footprint(elements_nb: int) -> None:
list_ = []
tuple_ = ()
print_list_tuple_comparison(list_, tuple_)
for i in range(1, elements_nb + 1):
list_.append(i)
tuple_ = tuple(range(1, i + 1))
print_list_tuple_comparison(list_, tuple_)
print_list_tuple_ram_footprint(10)Number of elements: 0, list size: 56 , tuple size: 40 , diff: 16 bytes (28.57%) Number of elements: 1, list size: 88 , tuple size: 48 , diff: 40 bytes (45.45%) Number of elements: 2, list size: 88 , tuple size: 56 , diff: 32 bytes (36.36%) Number of elements: 3, list size: 88 , tuple size: 64 , diff: 24 bytes (27.27%) Number of elements: 4, list size: 88 , tuple size: 72 , diff: 16 bytes (18.18%) Number of elements: 5, list size: 120, tuple size: 80 , diff: 40 bytes (33.33%) Number of elements: 6, list size: 120, tuple size: 88 , diff: 32 bytes (26.67%) Number of elements: 7, list size: 120, tuple size: 96 , diff: 24 bytes (20.00%) Number of elements: 8, list size: 120, tuple size: 104, diff: 16 bytes (13.33%) Number of elements: 9, list size: 184, tuple size: 112, diff: 72 bytes (39.13%) Number of elements: 10, list size: 184, tuple size: 120, diff: 64 bytes (34.78%)
Temps de création
Le tuple est plus rapide à créé que la list:
from timeit import timeit
tuple_elapsed_time = timeit("(1,2,3,4,5,6,7,8,9,10)", number=10_000_000)
list_elapsed_time = timeit("[1,2,3,4,5,6,7,8,9,10]", number=10_000_000)
print(f'Time needed to create a list of 10 ints 10_000_000 times: {list_elapsed_time}')
print(f'Time needed to create a tuple of 10 ints 10_000_000 times: {tuple_elapsed_time} ({(list_elapsed_time - tuple_elapsed_time) / list_elapsed_time * 100:.2f}% faster)')Time needed to create a list of 10 ints 10_000_000 times: 0.6825288010004442 Time needed to create a tuple of 10 ints 10_000_000 times: 0.0716032920172438 (89.51% faster)
Création d'une liste et accès à un de ses membres
from dis import dis
def my_list():
x = [10, 20, 30, 'abc']
y = x[0]
dis(my_list) 4 0 LOAD_CONST 1 (10)
2 LOAD_CONST 2 (20)
4 LOAD_CONST 3 (30)
6 LOAD_CONST 4 ('abc')
8 BUILD_LIST 4
10 STORE_FAST 0 (x)
5 12 LOAD_FAST 0 (x)
14 LOAD_CONST 5 (0)
16 BINARY_SUBSCR
18 STORE_FAST 1 (y)
20 LOAD_CONST 0 (None)
22 RETURN_VALUE
Création d'un tuple et accès à un de ses membres
from dis import dis
def my_tuple():
x = (10, 20, 30, 'abc')
y = x[0]
dis(my_tuple) 4 0 LOAD_CONST 1 ((10, 20, 30, 'abc'))
2 STORE_FAST 0 (x)
5 4 LOAD_FAST 0 (x)
6 LOAD_CONST 2 (0)
8 BINARY_SUBSCR
10 STORE_FAST 1 (y)
12 LOAD_CONST 0 (None)
14 RETURN_VALUE
La création d'un tuple est plus rapide du fait de la Python peephole optimization.