this I developed and it has helped me out
//how to return results - added in v1.34
define('XML_RETURN_ALL',1);
define('XML_RETURN_NEST',2);
define('XML_RETURN_FIRST',3);
//what parameters to return - added in v1.34
define('XML_PARAMS_POSN',1);
define('XML_PARAMS_ADDMID',2);
$xml_read_tags['version']=1.34;
function xml_read_tags($str,$tag, $return=XML_RETURN_ALL, $params=XML_PARAMS_POSN, $case=1){
//this function returns all instances of a given tag in a string, indexed 1-based, with an array structure for all of the attributes. Here is the array structure:
/*----------------------------------------------------------
2005-10-30: version 1.34 - fixed a bug with closing tag, streamlined coding, added the ability to get the "mid" string of a tag and also to return just the first instance if a tag is known to be unique. Two changes in output also:
1. instead of ['_special'] we now return the integer [0]
2. instead of 'int_strposn' and 'int_length' we now return 'mid_..' which is more understandable
2004-07-11: added case=1, all attribute names are lcase, 2=all are uppercase
2004-03-20: added flexiblity in the name call, can search for tags via regex, e.g. passing xml_read_tags($str, '((list)|(focus))*Element') will find: 1) listElement, 2)focusElement, and 3) element, another example would be -- t(r)|(head) will find either tr or thead in a table structure.
2004-03-12: added 'int_'strposn'' (intermediate text position) and 'int_length' (intermediate text length) for ease of use
Array Structure returned:
1 => attribute1 => value1
attribute2 => value2
..
0 => name => tag_name
0 => length => n1
0 => 'strposn' => n2
..
[no other 0 keys at this time]
2 => same as instance one
----------------------------------------------------------*/
global $xml_read_tags;
$regex='/(<('.$tag.')((\s+[^>]+>)|(>)))(.|\s)*$/i';
$mstLen=strlen($str);
$mstStr=$str;
while(true){
$fs++;
if($fs>200)break;
ob_start();
$ok=preg_match($regex,$str,$right);
$x=ob_get_contents();
ob_end_clean();
if($x){$xml_read_tags[err]='Parse or other error reading the tag against the string';}
if(!$ok)break;
$i++;
if($return==XML_RETURN_FIRST && $i==2) return $a[1];
$offFrRt=strlen($right[0]);
$offFrLt=$mstLen-$offFrRt;
//here is the entire tag
$thisTag=$right[1];
$tagLen=strlen($thisTag);
$a[$i][0]['name']=$right[2];
$p1=($a[$i][0]['strposn']=$offFrLt);
$p2=($a[$i][0]['length']=$tagLen);
if(preg_match_all('/[a-z0-9_-]+\s*((=\s*\'[^\']*\')|(=\s*"[^"]*")|(=[^ \/>]+))/i',$thisTag,$att)){
$a[$i][0]['attributes']=count($att[0]);
//iterate through each attribute = $j
for($j=0;$j<count($att[0]);$j++){
#split by =, declare the array
$thisTagAtt=$att[0][$j];
$label=trim(substr($thisTagAtt,0,strpos($thisTagAtt,'=')));
$value=substr($thisTagAtt,strpos($thisTagAtt,'=')+1);
if(preg_match('/^\'(.|\s)*\'$/',$value)){
$value=preg_replace('/^\'/','',$value);
$value=preg_replace('/\'$/','',$value);
}
if(preg_match('/^"(.|\s)*"$/',$value)){
$value=preg_replace('/^"/','',$value);
$value=preg_replace('/"$/','',$value);
}
#set the values in array
$case==1?$label=strtolower($label):'';
$case==2?$label=strtoupper($label):'';
$a[$i][$label]=$value;
}
}
//munch the name value pairs and look for flags like nowrap
$munch = preg_replace('/[a-z0-9_-]+\s*((=\s*\'[^\']*\')|(=\s*"[^"]*")|(=[^ ]+))(\s|(\/>)|>)/i','',$thisTag);
$munch= preg_replace('/<'.$tag.'/i','',$munch);
$munch= preg_replace('/\/*>/','',$munch);
$munchlist=preg_split('/\s+/',trim($munch));
if(count($munchlist)){
foreach($munchlist as $u){
if($u!==''){
$case==1?$u=strtolower($u):'';
$case==2?$u=strtoupper($u):'';
$a[$i][$u]=1;
$a[$i][0][flags][]=$u;
}
}
}
//redeclare the string from the new position
$str=
($offFrRt>$tagLen?substr($str,-($offFrRt-strlen($thisTag))):'');
if(substr($mstStr,$p1+$p2-2,1)!=='/'){
//this is a nesting tag, three cases:
#1. we have a nesting tag, with no subnested similar tags
#2. we have a nesting tag with a subnested similar tag
#3. we have no closing tag -- bad XML form
unset($l);
unset($level);
unset($nxtStr);
$l=1;
$level[$l]=1;
$nxtStr=$str;
//========================================
while(true){
//fail safe
$fs2++; if($fs2>200)break;
//match an open or close tag downstream
$nxtStr=substr($str,-(strlen($nxtStr)-2));
preg_match('/<(\/*)'.$tag.'(.|\s)*$/i',$nxtStr,$match2);
$l++;
//get the length of the string
$len=strlen($match2[0]);
if(substr($match2[0],1,1)=='/'){
$level[$l]=-1;
}else{
$level[$l]=1;
}
if(array_sum($level)==0){
// the end tag starts $len from the end of the string
$p3=$a[$i][0]['end_strposn']=$mstLen-$len;
#this is risky - though 99.99% of closing tags will be in format </tag>
$p4=$a[$i][0]['end_length']=strlen($a[$i][0]['name'])+3;
$p5=$a[$i][0]['total_length']=$a[$i][0]['end_strposn'] - $a[$i][0]['strposn']+$a[$i][0]['end_length'];
$a[$i][0]['mid_strposn']=$p1+$p2;
$a[$i][0]['mid_length']=($p3-$p1)-$p2;
if($params==XML_PARAMS_ADDMID) $a[$i][0]['mid_content']=substr($mstStr,$p1+$p2,$p3-$p1-$p2);
break;
}else{
$nxtStr=substr($nxtStr, -($len));
}
}
//========================================
}
}
return ($return==XML_RETURN_FIRST ? $a[1] : $a);
}
