XML 解析器函数
在线手册：中文 英文
PHP手册

xml_set_element_handler

(PHP 4, PHP 5)

xml_set_element_handler — 建立起始和终止元素处理器

说明

bool xml_set_element_handler ( resource $parser , callback $start_element_handler , callback $end_element_handler )

为 parser 参数指定的 XML 解析器建立元素处理器函数。参数 start_element_handler 和 end_element_handler 为表示函数名称的字符串，这些函数必须在为 parser 指定的解析器调用 xml_parse() 函数时已存在。

由 start_element_handler 参数命名的函数名必须接受三个参数：

start_element_handler ( resource $parser , string $name , array $attribs )
parser
第一个参数 parser 为指向要调用处理器的 XML 解析器的指针。
name
第二个参数 name 为该处理器为之被调用的元素名。如果大小写折叠（case-folding）对该解析器有效，元素名将用大写字母表示。
attribs
第三个参数 attribs 为一个包含有对应元素的属性的数组（如果该元素有属性）。数组元素的下标为属性名，元素的值即为属性的值。属性名将以和元素名同样的标准进行大小写折叠（case-folded），其值不进行大小写折叠。 属性的原始顺序将会被参数保留，用 each() 函数遍历 attribs 时，该数组下表的顺序和属性的顺序相同。

由 end_element_handler 参数命名的函数名必须接受两个参数：

end_element_handler ( resource $parser , string $name )
parser
第一个参数 parser 为指向要调用处理器的 XML 解析器的指针。
name
第二个参数 name 为该处理器为之被调用的元素名。如果大小写折叠（case-folding）对该解析器有效，元素名将用大写字母表示。

如果处理器函数名被设置为空字符串或者 FALSE，则该有问题的处理器将被屏蔽。

如果处理器被成功的建立，该函数将返回 TRUE；如果 parser 指向的不是合法的解析器，函数该函数将返回 FALSE。

Note: 除了函数名，含有对象引用的数组和方法名也可以作为参数。

参数

parser

A reference to the XML parser to set up start and end element handler functions.

start_element_handler

The function named by start_element_handler must accept three parameters:

start_element_handler ( resource $parser , string $name , array $attribs )
parser
The first parameter, parser, is a reference to the XML parser calling the handler.
name
The second parameter, name, contains the name of the element for which this handler is called.If case-folding is in effect for this parser, the element name will be in uppercase letters.
attribs
The third parameter, attribs, contains an associative array with the element's attributes (if any).The keys of this array are the attribute names, the values are the attribute values.Attribute names are case-folded on the same criteria as element names.Attribute values are not case-folded. The original order of the attributes can be retrieved by walking through attribs the normal way, using each().The first key in the array was the first attribute, and so on.

Note: 除了函数名，含有对象引用的数组和方法名也可以作为参数。

end_element_handler

The function named by end_element_handler must accept two parameters:

end_element_handler ( resource $parser , string $name )
parser
The first parameter, parser, is a reference to the XML parser calling the handler.
name
The second parameter, name, contains the name of the element for which this handler is called.If case-folding is in effect for this parser, the element name will be in uppercase letters.

If a handler function is set to an empty string, or FALSE, the handler in question is disabled.

返回值

成功时返回 TRUE， 或者在失败时返回 FALSE.

XML 解析器函数
在线手册：中文 英文
PHP手册
PHP手册 - N: 建立起始和终止元素处理器

用户评论:

kok at nachon dot nl (26-Sep-2008 02:44)

Here is another example of detecting empty elements. I works with libxml2. Note that it handles buffer boundaries.

<?php

$depth = 0; //current depth, used for pretty printing
$empty = false; //whether the tag is empty
$offset = 0; //the index of the start of the current buffer within the stream

function tagStart($parser, $name, array $attribs) {
    global $depth, $empty, $data, $offset, $lastchar;
    $idx = xml_get_current_byte_index($parser);
    /* xml_get_current_byte_index returns index within the streams and not
       within the buffer.*/
   
    /* Check if the index is within the buffer. */
    if (isset($data[$idx - $offset])) {
        $c = $data[$idx - $offset];
    } else {
        /* If it isn't simple use the last character of the buffer. */
        $c = $lastchar;
    }
    $empty = $c == '/';
    echo str_repeat("\t", $depth), "<$name", ($empty ? '/>' : '>'), "\n";
    if (!$empty) ++$depth;
}

function tagEnd($parser, $name) {
    global $depth, $empty;
    if (!$empty) {
        --$depth;
        echo str_repeat("\t", $depth), "</$name>\n";
    } else {
        $empty = false;
    }
}

$parser = xml_parser_create();
xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, false);
xml_set_element_handler($parser, 'tagStart', 'tagEnd');

$data1 = '
<test>
    <empty att="3" />
    <nocontent></nocontent>
    <content>
        <empty/>
        <empty/>
    </content>
    <empty/';

$data2 = '>
    <empty att="5" />
</test>
';

$data = &$data1;
$length = strlen($data1);
$lastchar = $data[$length-1];
xml_parse($parser, $data1);
$offset .= $length;
$data = &$data2;
xml_parse($parser, $data2);

darien at etelos dot com (12-Feb-2007 06:48)

This documentation is somewhat awry. I know it's been said many times before, but it bears repeating...

If using PHP4, you may be required to use xml_set_object() instead of calling any of the xml_set_*_handler() functions with a two-item array. It will work fine on PHP5, but move the same code to PHP4 and it will create one copie of $this (even if you use &$this) for each handler you set!

<?php
// This code will fail mysteriously on PHP4.
$this->parser = xml_parser_create();
xml_set_element_handler(
            $this->parser,
            array(&$this,"start_tag"),
            array(&$this,"end_tag")
        );
        xml_set_character_data_handler(
            $this->parser,
            array(&$this,"tag_data")
        );
?>

<?php
// This code will work on PHP4.
$this->parser = xml_parser_create();
xml_set_object($this->parser,&$this);
xml_set_element_handler(
            $this->parser,
            "start_tag",
            "end_tag"
        );
        xml_set_character_data_handler(
            $this->parser,
            "tag_data"
        );
?>

lloeki at gmail dot com (31-Oct-2006 11:07)

I modified the previous script, so that it is associative. I find it more useful that way. BTW I prefer strtolower() things, but that's not mandatory at all.

<?php

$file = "data.xml";
$depth = 0;
$tree = array();
$tree['name'] = "root";
$stack[] = &$tree;

function startElement($parser, $name, $attrs) {
   global $depth;
   global $stack;
   global $tree;
 
   $element = array();
   foreach ($attrs as $key => $value) {
       $element[strtolower($key)]=$value;
   }

   end($stack);
   $stack[key($stack)][strtolower($name)] = &$element;
   $stack[strtolower($name)] = &$element;
  
   $depth++;
}

function endElement($parser, $name) {
   global $depth;
   global $stack;

   array_pop($stack);
   $depth--;
}

$xml_parser = xml_parser_create();
xml_set_element_handler($xml_parser, "startElement", "endElement");
if (!($fp = fopen($file, "r"))) {
   die("could not open XML input");
}

while ($data = fread($fp, 4096)) {
   if (!xml_parse($xml_parser, $data, feof($fp))) {
       die(sprintf("XML error: %s at line %d",
                   xml_error_string(xml_get_error_code($xml_parser)),
                   xml_get_current_line_number($xml_parser)));
   }
}
xml_parser_free($xml_parser);
$tree = end(end($stack));
echo "<pre>";
print_r($tree);
echo "</pre>";

?>

rubentrancoso at gmail dot com (29-Aug-2006 09:44)

My 25 cents. This example show how to parse a XML in a associative array tree.

<?php

$file = "flow/flow.xml";
$depth = 0;
$tree = array();
$tree['name'] = "root";
$stack[count($stack)] = &$tree;

function startElement($parser, $name, $attrs) {
   global $depth;
   global $stack;
   global $tree;
  
   $element = array();
   $element['name'] = $name;
   foreach ($attrs as $key => $value) {
        //echo $key."=".$value;
        $element[$key]=$value;
    }

   $last = &$stack[count($stack)-1];
   $last[count($last)-1] = &$element;
   $stack[count($stack)] = &$element;

   $depth++;
}

function endElement($parser, $name) {
   global $depth;
   global $stack;

   array_pop($stack);
   $depth--;
}

$xml_parser = xml_parser_create();
xml_set_element_handler($xml_parser, "startElement", "endElement");
if (!($fp = fopen($file, "r"))) {
   die("could not open XML input");
}

while ($data = fread($fp, 4096)) {
   if (!xml_parse($xml_parser, $data, feof($fp))) {
       die(sprintf("XML error: %s at line %d",
                   xml_error_string(xml_get_error_code($xml_parser)),
                   xml_get_current_line_number($xml_parser)));
   }
}
xml_parser_free($xml_parser);
$tree = $stack[0][0];
echo "<pre>";
print_r($tree);
echo "</pre>";

redb (15-May-2006 01:48)

Example below (BadParser) works fine with some changes.

xml_set_element_handler ( $parser, array ( &$this, 'tagStart' ), array ( &$this, 'tagEnd' ) );
xml_set_character_data_handler ( $parser, array ( &$this, 'tagContent' ) );

Anonymous Koward (06-Apr-2006 06:38)

This has been mentioned before, but I just spent several days trying to figure out what was going on. Folks, if your XML parser is completely in a class, look at the documentation for xml_set_object(). The documentation above does say you can use functions inside classes as callbacks for xml_set_element_handler, but it doesn't tell you that if your entire XML parser is inside a class, then this is the _WRONG_ way to do things.

You should instead call xml_set_object with your parser variable and $this, which will then fix strange errors that can otherwise crop up and stop you having to pass an array of ( $this, 'tagStart' ) to this function.

e.g.

<?php

class BadParser
{

    function BadParser ()
    {

        $parser = xml_parser_create();

        //This is the WRONG WAY to set the functions inside your class for parsing the XML.
        xml_set_element_handler ( $parser, array ( $this, 'tagStart' ), array ( $this, 'tagEnd' ) );
        xml_set_character_data_handler ( $parser, array ( $this, 'tagContent' ) );

        xml_parse ( $parser, $this->XMLData );

    }

    function tagStart ( $parser, $tagName, $attributes = NULL )
    {

        $this->tag = $tagName;

    }

    function tagEnd ( $parser, $tagName )
    {

        $this->tag = NULL;

    }

    function tagContent ( $parser, $content )
    {

        //This WILL NOT work as you intended. $this->tag will do strange, mysterious things, but it won't be the tag name like you expected.
        echo ( "{$this->tag}: $content" );

    }

}

?>

Instead, you should change your constructor to do this as XML initalization instead:

<?php

class GoodParser
{

    function GoodParser ()
    {

        $parser = xml_parser_create();

        //This is the RIGHT WAY to set everything inside the object.
        xml_set_object ( $parser, $this );

        xml_set_element_handler ( $parser, 'tagStart', 'tagEnd' );
        xml_set_character_data_handler ( $parser, 'tagContent' );

        xml_parse ( $parser, $this->XMLData );

    }

    /* ... */

}

?>

I don't know if this problem exists in other versions of PHP. My version is 4.4.1. Hope I made sense, if this note had been around, it would've saved a lot of headaches for me (maybe I'm not observant enough).

vladimir-leontiev at uiowa dot edu (13-Dec-2005 05:44)

It seems that characterData() gets characters in chuncks of 1024; therefore if you have string of characters between you tags that is longer than 1024 then characterData() will be called more that once for single pair of tags. I don't know if this feature(bug?) is documented anywhere, I just wanted to warn everyone about this; it had tripped me. I use php 4.3.10 on Linux.

hendra_g at hotmail dot com (19-Nov-2005 04:10)

I ran into the same problem with 'ibjoel at hotmail dot com' in regards to self-closing tags, and found that the script that he/she wrote did not work as I expected.
I played around with some of php's functions and examples and compiled something, which may not be the neatest solution, but it works for the data that 'ibjoel at hotmail dot com' provided.
The data needs to be read from a file though, so the fp can be utilised. It still uses the xml_get_current_byte_index(resource parser) trick, but this time, I check for the last 2 character before the index and test if it's "/>".

<?php
/* myxmltest.xml:
<normal_tag>
  <self_close_tag />
     data
  <normal_tag>data
     <self_close_tag attr="value" />
  </normal_tag>
     data
  <normal_tag></normal_tag>
</normal_tag>
*/

//## Global Variables ##//
$file = "myxmltest.xml";
$character_data_on = false;
$tag_complete = true;

function startElement($parser, $name, $attrs)
{
    global $character_data_on;
    global $tag_complete;
   
    echo "&lt;<font color=\"#0000cc\">$name</font>";
    //## Print the attributes ##//
    if (sizeof($attrs)) {
        while (list($k, $v) = each($attrs)) {
            echo " <font color=\"#009900\">$k</font>=\"<font
                   color=\"#990000\">$v</font>\"";
        }
    }
    //## Tag is still still incomplete,
    //## will be completed at either endElement or characterData ##//
    $tag_complete = false;
    $character_data_on = false;
}

function endElement($parser, $name)
{
    global $fp;
    global $character_data_on;
    global $tag_complete;
   
    //#### Test for self-closing tag ####//
    //## xml_get_current_byte_index(resource parser) when run in this
    //## function, gives the index at (indicated by *):
    //##   for self closing tag: <br />*
    //##   for individual closing tag: <div>character data*</div>
    //## So to test for self-closing tag, we can just test for the last 2
    //## characters from the index
    //###################################//
   
    if (!$character_data_on) {
        //## Record current fp position ##//
        $temp_fp = ftell($fp);
       
        //## Point fp to 2 bytes before the end element byte index ##//
        $end_element_byte_index = xml_get_current_byte_index($parser);
        fseek($fp,$end_element_byte_index-2);
       
        //## Gets the last 2 characters before the end element byte index ##//
        $validator = fgets($fp, 3);
       
        //## Restore fp position ##//
        fseek($fp,$temp_fp);
       
        //## If the last 2 character is "/>" ##//
        if ($validator=="/>") {
            //// Complete the self-closing tag ////
            echo " /&gt";
            //// Otherwise it is an individual closing tag ////
        } else echo "&gt&lt/<font color=\"#0000cc\">$name</font>&gt";
        $tag_complete = true;
    } else echo "&lt/<font color=\"#0000cc\">$name</font>&gt";
   
    $character_data_on = false;
}

function characterData($parser, $data)
{
    global $character_data_on;
    global $tag_complete;
   
    if ((!$character_data_on)&&(!$tag_complete)) {
        echo "&gt";
        $tag_complete = true;
    }
    echo "<b>$data</b>";
    $character_data_on = true;
}

$xml_parser = xml_parser_create();
xml_parser_set_option($xml_parser, XML_OPTION_CASE_FOLDING, false);
xml_set_element_handler($xml_parser, "startElement", "endElement");
xml_set_character_data_handler($xml_parser, "characterData");
if (!($fp = fopen($file, "r"))) {
    die("could not open XML input");
}

echo "<pre>";
while ($file_content = fread($fp, 4096)) {
    if (!xml_parse($xml_parser, $file_content, feof($fp))) {
        die(sprintf("XML error: %s at line %d",
                    xml_error_string(xml_get_error_code($xml_parser)),
                    xml_get_current_line_number($xml_parser)));
    }
}
echo "</pre>";
xml_parser_free($xml_parser);
?>

turan dot yuksel at tcmb dot gov dot tr (20-Sep-2005 11:41)

The method that 'ibjoel at hotmail dot com' have described requires libxml2 as the xml parser, it does not work with expat. For a brief explanation, see xml_get_current_byte_index.

ibjoel at hotmail dot com (01-Aug-2005 11:49)

I noticed that in the example below, and all the examples I've seen on this site for viewing xml in html, the look of self closing tags such as <br /> are not preserved. The parser cannot distinguish between <tag /> and <tag></tag>, and if your start and end element functions are like these examples, both instances will be output with both an indvidual start and end tag.  I needed to preserve self-closing tags and it took me a while to figure out this work around. Hope this helps someone...
 
The start tag is left open, and then completed by it's first child, the next start tag or its end tag.  The end tag will complete with " />", or </tag> depending on the number of bytes between the start and end tags in the parsed data.
<?php
//$data=filepath or string
$data=<<<DATA
<normal_tag>
  <self_close_tag />
      data
  <normal_tag>data
     <self_close_tag attr="value" />
  </normal_tag>
      data
  <normal_tag></normal_tag>
</normal_tag>
DATA;

function startElement($parser, $name, $attrs)
{
        xml_set_character_data_handler($parser, "characterData");
        global $first_child, $start_byte;
        if($first_child)          //close start tag if neccessary
                echo "><br />";
        $first_child=true;
        $start_byte=xml_get_current_byte_index ($parser);
        if(count($attrs)>=1){
                foreach($attrs as $x=>$y){
                        $attr_string .= " $x=\"$y\"";
                }
        }
        echo htmlentities("<{$name}{$attr_string}"); //unclosed starttag
}

function endElement($parser, $name)
{
        global $first_child, $start_byte;
        $byte=xml_get_current_byte_index ($parser);
        if($byte-$start_byte>2){           //if end tag is more than 2 bytes from start tag
                if($first_child)          //close start tag if neccessary
                        echo "><br />";
                echo htmlentities("</{$name}>")."<br />";  //individual end tag
        }else
                echo " /><br />";  // self closing tag
        $first_child=false;

}

function characterData($parser, $data)
{
        global $first_child;
        if($first_child)  //if $data is first child, close start tag
                echo "><br />";
        if($data=trim($data))
                echo "<font color='blue'>$data</font><br />";
        $first_child=false;
}

function ParseData($data)
{
        $xml_parser = xml_parser_create();
        xml_set_element_handler($xml_parser, "startElement", "endElement");
        xml_parser_set_option($xml_parser,XML_OPTION_CASE_FOLDING,0);
        if(is_file($data))
        {
                if (!($fp = fopen($file, "r"))) {
                        die("could not open XML input");
                }

                while ($data = fread($fp, 4096)) {
                        if (!xml_parse($xml_parser, $data, feof($fp))) {
                                $error=xml_error_string(xml_get_error_code($xml_parser));
                               $line=xml_get_current_line_number($xml_parser);
                                die(sprintf("XML error: %s at line %d",$error,$line));
                        }
                }
        }else{
                if (!xml_parse($xml_parser, $data, 1)) {
                                $error=xml_error_string(xml_get_error_code($xml_parser));
                                $line=xml_get_current_line_number($xml_parser);
                                die(sprintf("XML error: %s at line %d",$error,$line));
                }
        }
       
        xml_parser_free($xml_parser);
}

ParseData($data);
?>

youniforever at naver dot com (30-Apr-2005 11:41)

<html>
  <head>
    <title>SAX Demonstration</title>
   <META HTTP-EQUIV='Content-type' CONTENT='text/html; charset=euc-kr'>
  </head>
  <body>
    <h1>RSS </h1>
   
      <?php

     $file = "data.xml";
     
      $currentTag = "";
      $currentAttribs = "";

      function startElement($parser, $name, $attribs)
      {
          global $currentTag, $currentAttribs;
          $currentTag = $name;
   
          $currentAttribs = $attribs;
          switch ($name) {
         
          default:
              echo("<b>&lt$name&gt</b><br>");
              break;
          }
      }

      function endElement($parser, $name)
      {
          global $currentTag;
          switch ($name) {
          default:
              echo("<br><b>&lt/$name&gt</b><br><br>");
              break;
          }
          $currentTag = "";
          $currentAttribs = "";
      }

      function characterData($parser, $data)
      {
          global $currentTag;
          switch ($currentTag) {
          case "link":
              echo("<a href=\"$data\">$data</a>\n");
              break;
          case "title":
              echo("title : $data");
              break;
          default:
              echo($data);
              break;
          }
      }

     $xmlParser = xml_parser_create();
   
      $caseFold = xml_parser_get_option($xmlParser,
                                        XML_OPTION_CASE_FOLDING);
   
      $targetEncoding = xml_parser_get_option($xmlParser,
                                              XML_OPTION_TARGET_ENCODING);

      if ($caseFold == 1) {
          xml_parser_set_option($xmlParser, XML_OPTION_CASE_FOLDING, false);
      }

      xml_set_element_handler($xmlParser, "startElement", "endElement");
      xml_set_character_data_handler($xmlParser, "characterData");

      if (!($fp = fopen($file, "r"))) {
          die("Cannot open XML data file: $file");
      }

     while ($data = fread($fp, 4096)) {
          if (!xml_parse($xmlParser, $data, feof($fp))) {
              die(sprintf("XML error: %s at line %d",
                          xml_error_string(xml_get_error_code($xmlParser)),
                          xml_get_current_line_number($xmlParser)));
              xml_parser_free($xmlParser);
          }
      }
      xml_parser_free($xmlParser);
      ?>
    </table>
  </body>
</html>

(13-Mar-2005 09:34)

In response to aw at avatartechnology dot com...
In response to landb at mail dot net...

When your functions are in an object:
Careful ! Don't forget to add: & (reference) to your parameters.

xml_set_element_handler($parser, array(&$this,"_startElement"), array(&$this,"_endElement"));
--> xmlparse will work on your object (good).

instead of:
xml_set_element_handler($parser, array($this,"_startElement"), array($this,"_endElement"));
---> xmlparse will work on a COPY of your object (often bad)

Vin-s
(sorry for my english)

aw at avatartechnology dot com (21-Oct-2004 04:40)

In response to landb at mail dot net...

As the notes mention, you can pass an array that contains the reference to an object and a method name when you need... so you can call methods in your own class as handlers like this:

xml_set_element_handler($parser, array($this,"_startElement"), array($this,"_endElement"));

Hope it helps...

tj at tobyjoe dot com (31-Jan-2003 07:22)

It seems that the tag handlers don't block on one another (the end handler is called whether or not the begin handler has finished). this can put you in a tight spot if you don't realize it while planning your app.

(11-Oct-2001 12:09)

You CAN use classes to parse XML. Just take a look at the following function:

xml_set_object

jg at jmkg dot net (07-May-2001 07:08)

If you are using a class for xml parsing, and want to check the return value of xml_set_element_handler in case it  fails, you must do this outside of the class's constructor. Inside the constructor, PHP-4.0.5 will die.

Basically, put all your xml initialisation code in another function of the class, and keep it out of the constructor.