PHP采集程序常用的采集函数收藏

获取所有链接内容和地址
function getAllURL($code){
preg_match_all('/<as+href=["|']?([^>"' ]+)["|']?s*[^>]*>([^>]+)</a>/i',$code,$arr);
return array('name'=>$arr[2],'url'=>$arr[1]);
}

获取所有的图片地址
function getImgSrc($code){
$reg = "/]*src="(http://(.+)/(.+).(jpg|gif|bmp|bnp|png))"/isU";
preg_match_all($reg, $code, $img_array, PREG_PATTERN_ORDER);
return $img_array[1];
}

当前的脚本网址
function getSelfURL(){
if(!empty($_SERVER["REQUEST_URI"])){
$scriptName = $_SERVER["REQUEST_URI"];
$nowurl = $scriptName;
}else{
$scriptName = $_SERVER["PHP_SELF"];
if(empty($_SERVER["QUERY_STRING"])) $nowurl = $scriptName;
else $nowurl = $scriptName."?".$_SERVER["QUERY_STRING"];
}
return $nowurl;
}

把全角数字转为半角数字
function getAlabNum($fnum){
$nums = array("０","１","２","３","４","５","６","７","８","９");
$fnums = "0123456789";
for($i=0;$i<=9;$i++) $fnum = str_replace($nums[$i],$fnums[$i],$fnum);
$fnum = ereg_replace("[^0-9.]|^0{1,}","",$fnum);
if($fnum=="") $fnum=0;
return $fnum;
}

去除HTML标记
function text2Html($txt){
$txt = str_replace(" ","　",$txt);
$txt = str_replace("<","<",$txt);
$txt = str_replace(">",">",$txt);
$txt = preg_replace("/[rn]{1,}/isU","<br/>rn",$txt);
return $txt;
}

清除HTML标记
function clearHtml($str){
$str = str_replace('<','<',$str);
$str = str_replace('>','>',$str);
return $str;
}

相对路径转化成绝对路径
function relative2Absolute($content, $feed_url) {
preg_match('/(http|https|ftp):///', $feed_url, $protocol);
$server_url = preg_replace("/(http|https|ftp|news):///", "", $feed_url);
$server_url = preg_replace("//.*/", "", $server_url);
if ($server_url == '') {
return $content;
}
if (isset($protocol[0])) {
$new_content = preg_replace('/href="//', 'href="'.$protocol[0].$server_url.'/', $content);
$new_content = preg_replace('/src="//', 'src="'.$protocol[0].$server_url.'/', $new_content);
} else {
$new_content = $content;
}
return $new_content;
}

获取指定标记中的内容
function getTagData($str, $start, $end){
if ( $start == '' || $end == '' ){
return;
}
$str = explode($start, $str);
$str = explode($end, $str[1]);
return $str[0];
}

HTML表格的每行转为CSV格式数组
function getTrArray($table) {
$table = preg_replace("'<td[^>]*?>'si",'"',$table);
$table = str_replace("</td>",'",',$table);
$table = str_replace("</tr>","{tr}",$table);
//去掉 HTML 标记
$table = preg_replace("'<[/!]*?[^<>]*?>'si","",$table);
//去掉空白字符
$table = preg_replace("'([rn])[s]+'","",$table);
$table = str_replace(" ","",$table);
$table = str_replace(" ","",$table);
$table = explode(",{tr}",$table);
array_pop($table);
return $table;
}

将HTML表格的每行每列转为数组，采集表格数据
function getTdArray($table) {
$table = preg_replace("'<table[^>]*?>'si","",$table);
$table = preg_replace("'<tr[^>]*?>'si","",$table);
$table = preg_replace("'<td[^>]*?>'si","",$table);
$table = str_replace("</tr>","{tr}",$table);
$table = str_replace("</td>","{td}",$table);
//去掉 HTML 标记
$table = preg_replace("'<[/!]*?[^<>]*?>'si","",$table);
//去掉空白字符
$table = preg_replace("'([rn])[s]+'","",$table);
$table = str_replace(" ","",$table);
$table = str_replace(" ","",$table);
$table = explode('{tr}', $table);
array_pop($table);
foreach ($table as $key=>$tr) {
$td = explode('{td}', $tr);
array_pop($td);
$td_array[] = $td;
}
return $td_array;
}

返回字符串中的所有单词 $distinct=true 去除重复
function splitEnStr($str,$distinct=true) {
preg_match_all('/([a-zA-Z]+)/',$str,$match);
if ($distinct == true) {
$match[1] = array_unique($match[1]);
}
sort($match[1]);
return $match[1];
}

/*二维数组按指定的键值排序*/
function array_sort($array,$keys,$type='asc'){
 if(!isset($array) || !is_array($array) || empty($array)){
  return '';
 }
 if(!isset($keys) || trim($keys)==''){
  return '';
 }
 if(!isset($type) || $type=='' || !in_array(strtolower($type),array('asc','desc'))){
  return '';
 }
 $keysvalue=array();
 foreach($array as $key=>$val){
  $val[$keys] = str_replace('-','',$val[$keys]);
  $val[$keys] = str_replace(' ','',$val[$keys]);
  $val[$keys] = str_replace(':','',$val[$keys]);
  $keysvalue[] =$val[$keys];
 }
 asort($keysvalue); //key值排序
 reset($keysvalue); //指针重新指向数组第一个
 foreach($keysvalue as $key=>$vals) {
  $keysort[] = $key;
 }
 $keysvalue = array();
 $count=count($keysort);
 if(strtolower($type) != 'asc'){
  for($i=$count-1; $i>=0; $i--) {
   $keysvalue[] = $array[$keysort[$i]];
  }
 }else{
  for($i=0; $i<$count; $i++){
   $keysvalue[] = $array[$keysort[$i]];
  }
 }
 return $keysvalue;
}

array_sort($a,'price');

$arr = array(0, 1, 2, 3, 4);
$index = 2;
unset($arr[$index]);
echo count($arr);

if (empty($arr[$index])) {
    echo "arr[$index] is empty";
}

print_r($arr);

输出为：
4
arr[2] is empty
Array([0] => 0 [1] => 1 [3] => 3 [4] =>

$arr = array(0, 1, 2, 3, 4);
$rtn = array_splice($arr, 2, 1);
echo count($arr);
print_r($arr);
print_r($rtn);

输出为：
4
Array
([0] => 0 [1] => 1 [2] => 3 [3] => 4)
Array([0] =>

1、
$arr = array(0, 1, 2, 3, 4);
array_splice（$arr, 0, 1);

2、
$arr = array(0, 1, 2, 3, 4);
$arr = array_splice（$arr, 1);

 $fruits = array("apple","banana","orange","pear"); 

$fruit = array_pop($fruits); 

 //$fruits = array("apple","banana","orange"); 

//$fruit = "pear";

<?php
    $array = ('a' => "abc", 'b' => "bcd",'c' =>"cde",'d' =>"def",'e'=>"");
    array_filter($array);
    echo "<pre>";
    print_r($array);
?>
结果：
Array (

     [a] => abc

     [b] => bcd

     [c] => cde

    [d] => def

   )

$array = array('1', '2', '3', '4', '5');

$del_value = 3;
unset($array[array_search($del_value , $array)]);//利用unset删除这个元素

print_r($array);

输出

array('1', '2', '4', '5');

function string2array($data) { 
if($data == '') return array();
 @eval("$array = $data;"); return $array;

}

* @param    array   $data       数组
* @param    bool    $isformdata 如果为0，则不使用new_stripslashes处理，可选参数，默认为1
* string  返回字符串，如果，data为空，则返回空  */

function array2string($data, $isformdata = 1) {
 if($data == '') return '';
 if($isformdata) $data = new_stripslashes($data);
 return addslashes(var_export($data, TRUE));
}