目錄
和前面講到的 python線程互斥鎖Lock 類似,當有多個進程Process同時讀寫同一個文件時,爲了避免數據讀寫產生異常,我們需要爲正在操作的進程加上互斥鎖,互斥鎖的原理不管是對線程threading還是對進程Process而言都是一樣。
一.線程互斥鎖和進程互斥鎖注意事項
1.創建線程互斥鎖
# 導入線程threading模塊
import threading
# 創建線程互斥鎖
mutex = threading.Lock()
2.創建進程互斥鎖
from multip# 導入進程模塊
from multiprocessing import Process,Lock
# 創建進程互斥鎖
mutex = Lock()
注意導入模塊的區別,不要混淆使用!
二.進程互斥鎖Lock函數介紹
acquire()
— 鎖定資源;
release()
— 釋放資源;
三.進程互斥鎖Lock使用
案例一:使用進程,但不使用互斥鎖
from multiprocessing import Lock, Process
import time
import random
import os
def foo(i, mutex):
print('%s: %s is running' % (i, os.getpid()))
time.sleep(random.random())
print('%s:%s is done' % (i, os.getpid()))
if __name__ == '__main__':
mutex = Lock()
for i in range(10):
process = Process(target=foo, args=(i, mutex))
process.start()
輸出結果:
0: 17008 is running
1: 5288 is running
2: 1228 is running
3: 9724 is running
4: 7520 is running
5: 10236 is running
3:9724 is done
6: 16452 is running
7: 13328 is running
0:17008 is done
8: 9356 is running
9: 16432 is running
8:9356 is done
2:1228 is done
5:10236 is done
9:16432 is done
7:13328 is done
4:7520 is done
6:16452 is done
1:5288 is done
重輸出的結果來看,多個進程同時在操作,如果是對同一個文件讀寫操作,很明顯已經亂套了,這並不是我們想要的;如果多進程在讀寫同一文件時想要保證數據安全,必然需要加上互斥鎖,例如下面這個demo;
案例二:進程互斥鎖的使用
from multiprocessing import Lock, Process
import time
import random
import os
def foo(i, mutex):
mutex.acquire()
print('%s: %s is running' % (i, os.getpid()))
time.sleep(random.random())
print('%s:%s is done' % (i, os.getpid()))
mutex.release()
if __name__ == '__main__':
mutex = Lock()
for i in range(10):
process = Process(target=foo, args=(i, mutex))
process.start()
輸出結果:
0: 6908 is running
0:6908 is done
1: 7976 is running
1:7976 is done
3: 7824 is running
3:7824 is done
2: 17328 is running
2:17328 is done
4: 7844 is running
4:7844 is done
5: 15900 is running
5:15900 is done
6: 12648 is running
6:12648 is done
7: 16516 is running
7:16516 is done
8: 17348 is running
8:17348 is done
9: 13180 is running
9:13180 is done
完美,即便是對同一個文件進行讀寫操作,進程Process使用互斥鎖Lock之後也不會造成數據混亂的問題,同時也提高了效率,完美解決案例一的問題!
案例三:對全局變量累計求和看看計算結果
# !usr/bin/env python
# -*- coding:utf-8 _*-
"""
@Author:何以解憂
@Blog(個人博客地址): shuopython.com
@WeChat Official Account(微信公衆號):猿說python
@Github:www.github.com
@File:python_process_lock.py
@Time:2019/12/31 21:25
@Motto:不積跬步無以至千里,不積小流無以成江海,程序人生的精彩需要堅持不懈地積累!
"""
# 導入進程模塊
from multiprocessing import Process,Lock
num = 0
def get_sum1():
global num # 聲明全局變量
for i in range(10000):
num = num +1
print("get_sum1:",num)
def get_sum2():
global num # 聲明全局變量
for i in range(10000):
num = num + 1
print("get_sum2:", num)
def main():
global num # 聲明全局變量
p1 = Process(target=get_sum1)
p1.start()
p2 = Process(target=get_sum2)
p2.start()
p1.join()
p2.join()
print("main:",num)
if __name__ == "__main__":
main()
print("main exit")
輸出結果:
get_sum1: 10000
get_sum2: 10000
main: 0
main exit
可能有小夥伴會覺得很納悶,main函數中得num值怎麼會是0,明明主進程/兩個子進程都用關鍵字 global 聲明瞭全局變量,即便沒有互斥鎖,也應該是一個小於20000的隨機數,在文章 python 進程Process與線程threading區別 中有詳細講解,同一進程的所有線程共享該進程的所有資源,進程與進程之間資源相互獨立,互不影響(類似深拷貝);
上面的程序有三個進程,這就意味着num變量實際上有三份資源,其中兩個進程對num分別做了10000次累計加1,所以每個子進程的值都是10000,主進程沒有對num任何操作,所以主進程num值爲0;
猜你喜歡:
4.python進程Process與線程threading的區別
轉載請註明:猿說Python » python 進程互斥鎖Lock
技術交流、商務合作請直接聯繫博主
掃碼或搜索:猿說python
猿說python
微信公衆號 掃一掃關注