PHP使用CURL实现多线程抓取网页

<?php 
$urls = array(  
 'http://.sina../',  
 'http://.sohu./',  
 'http://.163./' 
); // 设置要抓取的页面URL  
   
$save_to='/test.txt';  // 把抓取的代码写入该文件   
  
$st = fopen($save_to,"a");  
$mh = curl_multi_init();   
  
foreach ($urls as $i => $url) {  
 $conn[$i] = curl_init($url);  
 curl_set($conn[$i], CURLOPT_USERAGENT, "Mozilla/4.0 (patible; MSIE 7.0; Windows NT 6.0)");  
 curl_set($conn[$i], CURLOPT_HEADER ,0);  
 curl_set($conn[$i], CURLOPT_CONNECTTIMEOUT,60);  
 curl_set($conn[$i], CURLOPT_FILE,$st); // 设置将爬取的代码写入文件  
 curl_multi_add_handle ($mh,$conn[$i]);  
} // 初始化  
   
do {  
 curl_multi_exec($mh,$active);  
} while ($active); // 执行  
   
foreach ($urls as $i => $url) {  
 curl_multi_remove_handle($mh,$conn[$i]);  
 curl_close($conn[$i]);  
} // 结束清理  
   
curl_multi_close($mh);  
fclose($st); 
?> 

<?php 
$urls = array(  
 'http://.sina../',  
 'http://.sohu./',  
 'http://.163./' 
);  
  
$save_to='/test.txt';  // 把抓取的代码写入该文件  
$st = fopen($save_to,"a");  
  
$mh = curl_multi_init();  
foreach ($urls as $i => $url) {  
 $conn[$i] = curl_init($url);  
 curl_set($conn[$i], CURLOPT_USERAGENT, "Mozilla/4.0 (patible; MSIE 7.0; Windows NT 6.0)");  
 curl_set($conn[$i], CURLOPT_HEADER ,0);  
 curl_set($conn[$i], CURLOPT_CONNECTTIMEOUT,60);  
 curl_set($conn[$i],CURLOPT_RETURNTRANSFER,true); // 设置不将爬取代码写到浏览器，而是转化为字符串  
 curl_multi_add_handle ($mh,$conn[$i]);  
}  
  
do {  
 curl_multi_exec($mh,$active);  
} while ($active);  
   
foreach ($urls as $i => $url) {  
 $data = curl_multi_getcontent($conn[$i]); // 获得爬取的代码字符串  
 fwrite($st,$data); // 将字符串写入文件。，也可以不写入文件，比如存入数据库  
} // 获得数据变量，并写入文件  
  
foreach ($urls as $i => $url) {  
 curl_multi_remove_handle($mh,$conn[$i]);  
 curl_close($conn[$i]);  
}  
  
curl_multi_close($mh);  
fclose($st);  
?>