Apprendre asyncio.wait() and Task Control

Glissez pour afficher le menu

asyncio.gather() is convenient but opinionated – it waits for all coroutines and returns results in input order. asyncio.wait() gives you more control: you decide when to stop waiting and what to do with each task individually.

Basic Usage

asyncio.wait() takes a set of tasks (not coroutines) and returns two sets: done and pending.


              12345678910111213141516171819202122232425
            
import asyncio
import httpx
import nest_asyncio

nest_asyncio.apply()

# Fetching posts and processing them as they complete
async def fetch_post(client, post_id):
    url = f"https://jsonplaceholder.typicode.com/posts/{post_id}"
    response = await client.get(url)
    return response.json()["title"]

async def main():
    async with httpx.AsyncClient() as client:
        tasks = {
            asyncio.create_task(fetch_post(client, post_id))
            for post_id in [1, 2, 3]
        }

        done, pending = await asyncio.wait(tasks)

    for task in done:
        print(task.result())

asyncio.run(main())

Unlike gather(), results arrive in completion order – whichever task finishes first is processed first.

Return Conditions

The return_when parameter controls when wait() returns:

asyncio.ALL_COMPLETED (default): waits for all tasks to finish;
asyncio.FIRST_COMPLETED: returns as soon as any one task finishes;
asyncio.FIRST_EXCEPTION: returns as soon as any task raises an exception.


              1234567891011121314151617181920212223242526272829303132
            
import asyncio
import httpx
import nest_asyncio

nest_asyncio.apply()

# Returning as soon as the first post is fetched
async def fetch_post(client, post_id):
    url = f"https://jsonplaceholder.typicode.com/posts/{post_id}"
    response = await client.get(url)
    return response.json()["title"]

async def main():
    async with httpx.AsyncClient() as client:
        tasks = {
            asyncio.create_task(fetch_post(client, post_id))
            for post_id in [1, 2, 3]
        }

        done, pending = await asyncio.wait(
            tasks,
            return_when=asyncio.FIRST_COMPLETED,
        )

    print(f"First result: {next(iter(done)).result()}")
    print(f"Still pending: {len(pending)} tasks")

    # Cancelling remaining tasks
    for task in pending:
        task.cancel()

asyncio.run(main())

Cancelling Tasks

Any task can be cancelled with task.cancel(). The task receives a CancelledError at its next await point.


              1234567891011121314151617181920212223242526
            
import asyncio
import nest_asyncio

nest_asyncio.apply()

async def slow_work(label):
    try:
        print(f"Starting {label}")
        await asyncio.sleep(5)  # Simulating slow I/O
        return f"Done {label}"
    except asyncio.CancelledError:
        print(f"{label} was cancelled")
        raise  # Re-raising is required

async def main():
    task = asyncio.create_task(slow_work("task_A"))

    await asyncio.sleep(1)  # Letting the task start
    task.cancel()

    try:
        await task
    except asyncio.CancelledError:
        print("Task cancelled successfully")

asyncio.run(main())

Always re-raise CancelledError inside a coroutine – swallowing it silently breaks cancellation propagation.

`gather()` vs `wait()`

Tout était clair ?

Merci pour vos commentaires !

Section 2. Chapitre 3

Demandez à l'IA

Posez n'importe quelle question ou essayez l'une des questions suggérées pour commencer notre discussion

Section 2. Chapitre 3

asyncio.wait() and Task Control

Basic Usage

Return Conditions

Cancelling Tasks

gather() vs wait()

`gather()` vs `wait()`