concurrent.futures.wait

Concurrency & Async

Meaning

Waits for a set of Future objects to complete, optionally with a timeout and a condition for when to return (e.g., when the first future completes). Returns two sets: done and not_done futures.

Primary Function

Concurrency control

Communicative Purpose

Synchronize multiple asynchronous operations and react when a subset finishes.

Pattern

concurrent.futures.wait(futures, timeout=timeout, return_when=return_when)

Core Structure

concurrent.futures.wait(..., timeout=..., return_when=...)

Função primária

Concurrency control

Propósito comunicativo

Synchronize multiple asynchronous operations and react when a subset finishes.

Situações de gatilho

Waiting for any of several parallel tasks to finish; implementing a timeout for a group of futures; collecting results as soon as the first task completes.

Contextos

Python programs using concurrent.futures.ThreadPoolExecutor or ProcessPoolExecutor; async-like patterns with futures.

Padrão

concurrent.futures.wait(futures, timeout=timeout, return_when=return_when)

Estrutura central

concurrent.futures.wait(..., timeout=..., return_when=...)

Slots de substituição

futures: iterable of Future, timeout: float or None, return_when: constant from concurrent.futures (FIRST_COMPLETED, FIRST_EXCEPTION, ALL_COMPLETED)

Colocados típicos

ThreadPoolExecutor
ProcessPoolExecutor
Future.result
as_completed

Substituições comuns

using concurrent.futures.as_completed for iterative processing
using wait with timeout=None

Erros comuns

forgetting to handle the returned done/not_done sets; assuming wait returns results directly; using timeout incorrectly leading to premature return

Similar / contraste

concurrent.futures.as_completed (yields futures as they complete) vs wait (blocks until condition); threading.Event (simple barrier)

Interferências

Coming from Java's ExecutorService.invokeAny/invokeAll: may expect similar behavior but Python's wait returns sets, not results.

Família do chunk

concurrent.futures.as_completed
concurrent.futures.map
threading.Barrier

Nuance

wait does not cancel unfinished futures; timeout only affects the wait call, not the futures themselves; return_when determines which set is returned as done.

Efeito pragmático

Enables clear synchronization point for concurrent tasks, reducing boilerplate polling.

Dica de memória

Wait for the first future to cross the finish line.

Nota

wait() returns two sets of Futures (done, not_done); it does not cancel unfinished futures, and timeout only affects the wait call itself.

Upgrade path

Using asyncio.wait for coroutine-based concurrency

Frequência: MediumFormulaicidade: Semi-fixedTipo de construção: function callPrioridade de aquisição: Active recallPrioridade de output: BothTag de espaçamento: Medium-term

Examples

from concurrent.futures import ThreadPoolExecutor, wait, FIRST_COMPLETED with ThreadPoolExecutor() as exe: futures = [exe.submit(lambda: sum(range(i))) for i in range(10, 15)] done, not_done = wait(futures, timeout=2, return_when=FIRST_COMPLETED) print('Done:', [f.result() for f in done])

from concurrent.futures import ProcessPoolExecutor, wait, ALL_COMPLETED with ProcessPoolExecutor() as exe: futures = [exe.submit(pow, i, 2) for i in range(5)] done, not_done = wait(futures, timeout=None, return_when=ALL_COMPLETED) print([f.result() for f in done])

from concurrent.futures import ThreadPoolExecutor, wait, FIRST_EXCEPTION with ThreadPoolExecutor() as exe: futures = [exe.submit(lambda x: 1/x, i) for i in range(-2, 3)] done, not_done = wait(futures, return_when=FIRST_EXCEPTION) for f in done: if f.exception(): print('First exception:', f.exception()) break