Python 処理速度向上 Tips

実行環境

MacBook Pro
python3.7.5

ループ

while i < N　vs　for _ in range(N)

i = 0
while i < N:  # N=10**6
    i += 1

for _ in range(N):
    pass

1次元配列

配列の初期化

[None for _ in range(N)]　vs　[None] * N　vs　np.empty(N)

tmp = [None for _ in range(N)]

tmp = [None] * N

tmp = np.empty(N)

配列の要素が全て数値で良いなら， Numpyを用いた方が速い

要素の変換

内包表記　vs　map関数

tmp = [int(i) for i in range(N)]

tmp = list(map(int,range(N)))

2次元配列(N×N)

[None for_in range(N)]for_in range(N)] vs [[None]*N for_in range(N)] vs np.empty((N,N))

tmp = [None for _ in range(N]for _ in range(N)]

tmp = [[None] * N for _ in range(N)]

tmp = np.empty((N,N))