Be Careful Using Python Iterator

In a recent experiment, I encountered a bug where all the results were shown as negative. After debugging, I found that the bug was caused by using the same iterator twice. My library API was designed to take a list of tuples as input, but when I called it later, I passed in a map object.

My API extracts the first and second elements of the tuple and uses them independently to obtain the experiment results. However, every time I ran the experiment, the list of second elements was empty, resulting in all negative results.

So, the key takeaway is that a Python iterator can only be consumed once!!

Reproduction

def foo(x: int) -> tuple[int, int]:
    return (x, x + 1)


xs = [1, 2, 3]

# case 1: list comprehension
ys = [foo(x) for x in xs]
ys_fst = [y[0] for y in ys]
ys_snd = [y[1] for y in ys]

# case 2: map
zs = map(foo, xs)
zs_fst = [z[0] for z in zs]
zs_snd = [z[1] for z in zs]

print(ys_fst == zs_fst)
print(ys_snd == zs_snd)