前面说过由于GIL的存在,Python的多线程效率没有希望的那么高,python的多线程适合IO密集型的情况,而爬虫恰好就是一个IO密集的情况,因为爬虫中很大一部分时间,是在等待socket返回数据。
下面写一个例子:
import requests
import time if __name__ == '__main__':
codes = ['sh600993', 'sh000006', 'sh600658', 'sh600153', 'sh600005']
start = time.time()
for code in codes:
url = 'http://hq.sinajs.cn/list=' + code
response = requests.get(url).text
print response
print time.time() - start
var hq_str_sh600993="马应龙,20.020,20.090,20.060,20.060,19.950,20.040,20.060,486809,9740634.000,2100,20.040,8300,20.030,1300,20.020,2300,20.010,4100,20.000,101,20.060,10000,20.070,14400,20.080,19000,20.090,25700,20.100,2017-01-24,11:30:00,00"; var hq_str_sh000006="地产指数,6567.8364,6574.1060,6568.6375,6577.7249,6542.6599,0,0,1486830,1392918131,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2017-01-24,11:35:51,00"; var hq_str_sh600658="电子城,13.320,13.200,13.270,13.320,13.040,13.270,13.280,559733,7389992.000,30800,13.270,300,13.220,6200,13.200,2500,13.100,4900,13.090,9300,13.280,6400,13.290,8200,13.300,6900,13.310,9000,13.320,2017-01-24,11:30:00,00"; var hq_str_sh600153="建发股份,10.520,10.510,10.500,10.540,10.460,10.490,10.500,4834159,50730040.000,32800,10.490,60100,10.480,186000,10.470,181241,10.460,125800,10.450,56600,10.500,105500,10.510,108400,10.520,110400,10.530,139900,10.540,2017-01-24,11:30:00,00"; var hq_str_sh600005="武钢股份,0.000,3.710,3.710,0.000,0.000,0.000,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,2017-01-24,11:30:00,03"; 0.110999822617
换成多线程之后:
import requests
import threading
import time def get_stock(code):
url = 'http://hq.sinajs.cn/list=' + code
response = requests.get(url).text
# js_info = response.read()
print response if __name__ == '__main__':
codes = ['sh600993', 'sh000006', 'sh600658', 'sh600153', 'sh600005']
start = time.time()
threads = [threading.Thread(target = get_stock,args = (code,)) for code in codes]
for t in threads:
t.start()
for t in threads:
t.join()
print time.time()-start
var hq_str_sh600993="马应龙,20.020,20.090,20.060,20.060,19.950,20.040,20.060,486809,9740634.000,2100,20.040,8300,20.030,1300,20.020,2300,20.010,4100,20.000,101,20.060,10000,20.070,14400,20.080,19000,20.090,25700,20.100,2017-01-24,11:30:00,00"; var hq_str_sh600658="电子城,13.320,13.200,13.270,13.320,13.040,13.270,13.280,559733,7389992.000,30800,13.270,300,13.220,6200,13.200,2500,13.100,4900,13.090,9300,13.280,6400,13.290,8200,13.300,6900,13.310,9000,13.320,2017-01-24,11:30:00,00"; var hq_str_sh000006="地产指数,6567.8364,6574.1060,6568.6375,6577.7249,6542.6599,0,0,1486830,1392918131,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2017-01-24,11:35:51,00"; var hq_str_sh600153="建发股份,10.520,10.510,10.500,10.540,10.460,10.490,10.500,4834159,50730040.000,32800,10.490,60100,10.480,186000,10.470,181241,10.460,125800,10.450,56600,10.500,105500,10.510,108400,10.520,110400,10.530,139900,10.540,2017-01-24,11:30:00,00"; var hq_str_sh600005="武钢股份,0.000,3.710,3.710,0.000,0.000,0.000,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,0,0.000,2017-01-24,11:30:00,03"; 0.0379998683929
速度有了很大的提升
线程池
import requests
import threadpool
import time def get_stock(code):
url = 'http://hq.sinajs.cn/list=' + code
response = requests.get(url).text
# js_info = response.read()
print response if __name__ == '__main__':
codes = ['sh600993', 'sh000006', 'sh600658', 'sh600153', 'sh600005']
start = time.time()
pool = threadpool.ThreadPool(5)
tasks = threadpool.makeRequests(get_stock,codes)
[pool.putRequest(task) for task in tasks]
pool.wait()
print time.time() - start
threadpool.ThreadPool定义了一个线程池,表示可以创建4个线程;
makeRequests创建了要开启多线程的函数,已经函数的参数以及回调函数,回调函数callback可以不写,默认是无。
[pool.putRequest(task) for task in tasks]是将所有多线程的请求扔进了线程池,等价于
for code in codes:
pool.putRequest(code)
pool.wait()是等待所有工作完成后退出。这里执行的数量还比较少,基本的多线程就够用了,当数量多了起来之后,线程池的效果会好一些。