队列
python内置一个queue模块,包含四种队列
- queue.Queue: FIFO
- queue.LifoQueue: LIFO
- queue.PriorityQueue: 优先级队列
- queue.deque: 双向队列
队列的使用场景:
- 提高并发
- 流量削峰
- 程序解耦
基本生产者与消费者的Demo
from Queue import Queue as queue
def func(q):
# 生产者
for i in range(0, 10):
q.put(i)
# 消费者
for j in range(0, 10):
res = q.get(j)
print(j)
# 由消费者调用,每一个get()得到一个任务后,由task_done()通知队列此任务已经处理完成
q.task_done()
# 阻塞等待
q.join()
if __name__ == '__main__':
q = queue()
func(q)
多线程、队列
import threading
import time
import random
from Queue import Queue as queue
q = queue()
def func_put(index):
q.put(index)
def func_get(index):
q.get(index)
time.sleep(random.randint(1, 4))
print(index)
q.task_done()
def producer():
for i in range(0, 10):
t = threading.Thread(target = func_put, args = (i, ))
t.start()
def consumer():
for i in range(0, 10):
t = threading.Thread(target = func_get, args = (i, ))
t.start()
if __name__ == '__main__':
producer()
consumer()
q.join()
执行结果:
python demo-4.py
0
5
8
3
...
基本线程池(多线程+队列模型)
因为在切换线程时需要切换上下文环境,会造成cpu的大量开销,所以在使用多线程处理任务时并不是线程越多越好。为了解决这个问题,提出了线程池的逻辑模型。预先创建好一个较为优化的数量的线程,让过来的任务立刻能够使用,就形成了线程池。
import threading
import time
from Queue import Queue as queue
class MyThread():
def __init__(self, maxsize = 5):
self.maxsize = maxsize
self.q = queue(maxsize)
for i in range(maxsize):
self.q.put(threading.Thread)
def put_thread(self):
return self.q.put(threading.Thread)
def get_thread(self):
return self.q.get()
def task(index, pool):
time.sleep(1)
print(index)
pool.put_thread()
if __name__ == '__main__':
pool = MyThread(5)
for i in range(10):
t = pool.get_thread()
obj = t(target = task, args = (i, pool, ))
obj.start()
执行结果:
$ python demo-5.py
0
1
2
...