Duplicated results when spinning tasks with ThreadPoolExecutor.submit(lambda: x)

[hombit]

Bug report

Bug description:

Here’s the corrected version with Markdown syntax escaped where necessary:

import random
from concurrent.futures import ThreadPoolExecutor

random.seed(42)

data = [[random.random() for _ in range(1024)] for _ in range(16)]
with ThreadPoolExecutor(1) as pool:
futures = [pool.submit(lambda: sum(x)) for x in data]
results = [f.result() for f in futures]

print(results)

Output on CPython 3.13:

[523.3398190464421, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254, 509.139502449254]

If I run it differently — for example, with [pool.submit(sum, x) for x in data] or [pool.submit(partial(sum, x)) for x in data] — everything works as expected: I’m getting different results.

I can also reproduce it with a simpler version, using no sum or any other function, although it gives fewer duplicates, so the results are less stable:

import random
from concurrent.futures import ThreadPoolExecutor

random.seed(42)

data = [[random.random()] for _ in range(16)]
with ThreadPoolExecutor(1) as pool:
futures = [pool.submit(lambda: x) for x in data]
results = [f.result() for f in futures]

print(results)

I tested and reproduced the issue on the latest CPython releases from 3.9 to 3.13, and 3.14a4 on x86_64 Linux.
Additionally, I tested it on:

• aarch64 Linux CPython 3.13
• aarch64 macOS CPython 3.13
• x86_64 Linux PyPy 3.11

CPython versions tested on:

3.9, 3.10, 3.11, 3.12, 3.13, 3.14

Operating systems tested on:

Linux, macOS

[brianschubert]

Hi! This is https://stackoverflow.com/q/3431676/11082165. You're encountering the late-binding of x in the lambda expressions, not a CPython bug.

You can get the same behavior without ThreadPoolExecutor just by creating lambdas in a list comprehension:

funcs = [lambda: x for x in range(5)]
for f in funcs:
    print(f())

outputs:

4
4
4
4
4

[hombit]

Thank you, @brianschubert, closing it