php curl批处理--可控并发异步
在php的手册里面有一段代码：
$mh = curl_multi_init(); $ch = array(); $chunck = 10; //并发控制数 $all = count($urls);//所有的请求url数组 $chunck = $all > $chunck ? $chunck : $all;$options = array(CURLOPT_HEADER=>FALSE,CURLOPT_RETURNTRANSFER=>TRUE,CURLOPT_FOLLOWLOCATION=>TRUE,CURLOPT_MAXREDIRS=>5,CURLOPT_USERAGENT=>'Mozilla/5.0 (Windows NT 6.1; rv:6.0) Gecko/20100101 Firefox/6.0');for($i = 0 ; $i < $chunck ; $i++){$ch[$i] = curl_init();curl_setopt($ch[$i],CURLOPT_URL,$urls[$i]);curl_setopt_array($ch[$i],$options);curl_multi_add_handle($mh,$ch[$i]);}do { while(($execrun = curl_multi_exec($mh, $running)) == CURLM_CALL_MULTI_PERFORM); if($execrun != CURLM_OK)break; // a request was just completed -- find out which one while($done = curl_multi_info_read($mh)) { //获取已经返回的url在urls数组里德的index $index = array_search($done['handle'],$ch); $info = curl_getinfo($done['handle']); if ($info['http_code'] == 200){ $output = curl_multi_getcontent($ch[$index]); // request successful. process output using the callback function. $save_path = $i.'.txt';//数据保存路径file_put_contents($save_path,$output); // start a new request (it's important to do this before removing the old one) $next = $i++;// increment i $ch[$next] = curl_init(); $options[CURLOPT_URL] = $urls[$next];//将下一个请求添加到队列 curl_setopt_array($ch[$next],$options); curl_multi_add_handle($mh, $ch[$next]); // remove the curl handle that just completed curl_multi_remove_handle($mh, $done['handle']); } else { // request failed. add error handling. } } } while ($running);curl_multi_close($mh);