如何并行快速执行for循环（Python）-解网

问：

我正在使用 python 构建一个股票市场交易工具。截至目前，经纪公司发布了一个python sdk，我们可以使用它开发软件。当我发送购买股票的请求时，问题就出现了。每当下订单时，我都会收到来自服务器的响应作为确认，这需要 0.8 到 1.2 秒。

就我而言，按照上面的顺序，我一次至少会下 20 个订单，如果我使用 for 循环下订单，则至少需要 20 秒才能完成循环。实际上，我不使用在放置订单后获得的响应，因此我不想等待响应，并且需要一次执行所有 20 个订单。我尝试了线程、Asyncio、并发期货。所有这些循环至少需要 5 秒才能完成 20 次迭代的 for 循环。如果迭代小于 6，则使用线程会给出最佳结果，即订单会一直执行，没有任何延迟。我恳请有人对此有所了解，我怎样才能毫不拖延地执行 20 次迭代的 for 循环。

这需要一次执行 20 次：

client.place_order(exchange_segment = ExchangeSegment , product = Product, price = '45',
                   order_type = OrderType, quantity = quantity, validity = Validity,
                   trading_symbol = TradingSymbol, transaction_type = TransactionType,
                   tag = Tag, amo = AMO)

我尝试了以下方法，

线程：

def placeorder(quantity):
            client.place_order(exchange_segment = ExchangeSegment , product = Product, price = '45',order_type = OrderType, quantity = quantity, validity = Validity,trading_symbol = TradingSymbol, transaction_type = TransactionType,tag = Tag, amo = AMO)
            
        quantity=[]  
        for i in range(20):
          quantity.append(str(800)) #mentioning the order quantitiy for each iteration in this list
            
        
        threads = []
        for i in range(20):
            t = threading.Thread(target=palceorder, args=(quantity[i],))
            threads.append(t)
            t.start()

异步：

        quantity=[]  
        for i in range(20):
          quantity.append(str(800))
        def background(f):
            def wrapped(*args, **kwargs):
                return asyncio.get_event_loop().run_in_executor(None, f, *args, **kwargs)

            return wrapped

        @background
        def placeorder(quantity):
            client.place_order(exchange_segment = ExchangeSegment , product = Product, price = '45',order_type = OrderType, quantity = quantity, validity = Validity,trading_symbol = TradingSymbol, transaction_type = TransactionType,tag = Tag, amo = AMO)
        for i in quantity:
            placeorder(i)`

并发期货：

def placeorder(quantity):
            client.place_order(exchange_segment = ExchangeSegment , product = Product, price = '45',order_type = OrderType, quantity = quantity, validity = Validity,trading_symbol = TradingSymbol, transaction_type = TransactionType,tag = Tag, amo = AMO)
quantity=[]  
for i in range(20):
   quantity.append(str(800))
with concurrent.futures.ThreadPoolExecutor(15) as executor:
            executor.map(placeorder, quantity)

python-3.x 线程性能多处理

from collections import deque
from threading import Lock
import time
from concurrent.futures import ThreadPoolExecutor

class RateLimiter:
    def __init__(self, call_count: int, period: float=1.0) -> None:
        self._call_count = int(call_count)
        self._period = float(period)
        self._called_timestamps = deque()
        self._lock = Lock()

    def throttle(self) -> None:
        with self._lock:
            while True:
                now = time.monotonic()
                while self._called_timestamps:
                    time_left = self._called_timestamps[0] + self._period - now
                    if time_left >= 0:
                        break
                    self._called_timestamps.popleft()
                if len(self._called_timestamps) < self._call_count:
                    break
                time.sleep(time_left)
            self._called_timestamps.append(now)


# 20 calls per second (or perhaps 20 calls per 1.1 second
# just to be safe):
rate_limiter = RateLimiter(20, 1.0)

def create_order(x: int) -> None:
    ... # arbitrary code to get things ready for the API call

    # Put this immediately before invoking the API:
    rate_limiter.throttle()
    # We would normally be calling an API to create an order,
    # but for demo purposes we just print x with the time:
    print(f'x = {x:2} at time {time.time()}')

with ThreadPoolExecutor(20) as executor:
    executor.map(create_order, range(42))

指纹：

x =  0 at time 1700224533.0041072
x =  1 at time 1700224533.0041072
x =  2 at time 1700224533.0062273
x =  3 at time 1700224533.0062273
x =  5 at time 1700224533.0072248
x =  6 at time 1700224533.0082252
x =  4 at time 1700224533.0072248
x =  9 at time 1700224533.0082252
x = 10 at time 1700224533.0095797
x =  8 at time 1700224533.0082252
x = 13 at time 1700224533.0095797
x = 14 at time 1700224533.0105777
x = 11 at time 1700224533.0095797
x = 17 at time 1700224533.011574
x = 18 at time 1700224533.011574
x =  7 at time 1700224533.0082252
x = 16 at time 1700224533.011574
x = 12 at time 1700224533.0095797
x = 15 at time 1700224533.0105777
x = 19 at time 1700224533.011574
x = 20 at time 1700224534.011685
x = 21 at time 1700224534.011685
x = 23 at time 1700224534.011685
x = 24 at time 1700224534.011685
x = 26 at time 1700224534.011685
x = 28 at time 1700224534.011685
x = 22 at time 1700224534.011685
x = 31 at time 1700224534.011685
x = 33 at time 1700224534.011685
x = 25 at time 1700224534.011685
x = 29 at time 1700224534.011685
x = 32 at time 1700224534.011685
x = 27 at time 1700224534.011685
x = 30 at time 1700224534.011685
x = 34 at time 1700224534.026685
x = 35 at time 1700224534.026685
x = 37 at time 1700224534.026685
x = 39 at time 1700224534.026685
x = 38 at time 1700224534.026685
x = 36 at time 1700224534.026685
x = 40 at time 1700224535.025956
x = 41 at time 1700224535.025956

上一个：如何在 Amazon Sagemaker 笔记本（云）中设置 pytesseract 路径？

下一个：OSMnx 等时线 PolygonPatch IndexError：数组的索引过多：数组是 0 维的，但索引了 2 个

如何并行快速执行for循环（Python）

How to do rapid execution of for loop(Python) in parallel

评论