List Comprehension is better than Loop

List Comprehension	Loop
Syntax
`[i * i for i in range(10)]`	`result = [None]10 for i in range(10): result[i] = i i # append is slower`
Bytecode
`LOAD_FAST i LOAD_FAST i BINARY_MULTIPLY LIST_APPEND`	`LOAD_FAST i LOAD_FAST i BINARY_MULTIPLY LOAD_FAST result LOAD_FAST i STORE_SUBSCR`

Comparison Benchmark and Bytecodes

import dis
import timeit


def comprehension_version() -> list[int]:
    return [i * i for i in range(10)]


def loop_version() -> list[int]:
    result = [None] * 10
    for i in range(10):
        result[i] = i * i  # append is slower
    return result


comp_time = timeit.timeit(
    "comprehension_version()", globals=globals(), number=1_000_000
)
loop_time = timeit.timeit("loop_version()", globals=globals(), number=1_000_000)

print("Comprehension version time:", comp_time)  # 0.257528125
print("Loop version time:", loop_time)           # 0.27902224999999997

print("\nlist comprehension (outer) bytecodes:")
for instr in dis.Bytecode(comprehension_version):
    print(instr.opname, instr.argrepr)

# Extract the inner <listcomp> code object
inner_comp = [
    c
    for c in comprehension_version.__code__.co_consts
    if isinstance(c, type(comprehension_version.__code__))
][0]

print("\nlist comprehension (inner <listcomp>) bytecodes:")
for instr in dis.Bytecode(inner_comp):
    print(instr.opname, instr.argrepr)

print("\nfor loop bytecodes:")
for instr in dis.Bytecode(loop_version):
    print(instr.opname, instr.argrepr)

Last modified on 2026-01-26