PHP arrays, XML, and crap

[Zerodash]

I've got this nifty XML parser:

<?

class XmlParser {
var $params= array();
var $level = array();
	
function XmlParser($input) {
$xmlp = xml_parser_create();
xml_parse_into_struct($xmlp, $input, $vals, $index);
xml_parser_free($xmlp);
$this->updateMembers($vals, $index);
}
	
// Takes the data and puts it into PHP array  
function updateMembers($vals, $index) {
foreach ($vals as $xml_elem) {
  if ($xml_elem['type'] == 'open' ) {
    $this->level[$xml_elem['level']] = strtolower($xml_elem['tag']);
  }
  elseif ($xml_elem['type'] == 'complete') {
    $xml_elem['tag'] = strtolower($xml_elem['tag']);
    $start_level = 1;
    $php_stmt = '$this->params';
    while($start_level < $xml_elem['level']) {
      $php_stmt .= '[$this->level['.$start_level.']]';
      $start_level++;
    }
    $php_stmt .= '[$xml_elem[\'tag\']] = $xml_elem[\'value\'];';
    eval($php_stmt);
	
	
  }	
}
}
	
function getRoot() {
return $this->level[1];	
}
	
function getData() {
return $this->params;	
}
}
$xmlp = new XmlParser($xml_response);
$root = $xmlp->getRoot();
$data = $xmlp->getData();
?>

It works really well except for one problem- When parsing XML formatted like shown below, it only stores one "item" from the "item" element in the resulting array.

<order>
  <date>2006-09-19</date>
  <cart>
    <items>
      <item>
        <quantity>1</quantity>
        <unit-price currency=\"USD\">80.0</unit-price>
        <item-name>Garbage</item-name>
        <item-description>Smelly</item-description>
      </item>
      <item>
        <quantity>2</quantity>
        <unit-price currency=\"USD\">10.0</unit-price>
        <item-name>Another Item</item-name>
        <item-description>This item Sucks</item-description>
      </item>	  
      <item>
        <quantity>3</quantity>
        <unit-price currency=\"USD\">450.0</unit-price>
        <item-name>nachos</item-name>
        <item-description>Cheesey</item-description>
      </item>	  
    </items>
  </cart>
</order>

For instance, when I output echo $data[$root]['cart']['items']['item']['item-name'] . "<br>";, it will only echo "nachos". Presumably because as it loops through the parser, it copies over the "item" element over and over.

Does anyone have any idea what I can do to make sure all "items" are stored in this array? I cannot alter the XML in any way, so the parser needs to be fixed.

Thanks in advance for any input.

[cka]

the best way would be to have it as a numerically indexed array (or, even better, a mix-n-matched array so you can call $data['items'][n]['value'])

here's a better xml parser that'll pull off what you want it to do, it's pretty much the best xml->array parser I've ever seen that isn't 30 different files from PEAR:

<?php

class XMLParser {
    var $data;              // Input XML data buffer
    var $vals;              // Struct created by xml_parse_into_struct
    var $collapse_dups;     // If there is only one tag of a given name, shall we store as scalar or array?
    var $index_numeric;     // Index tags by numeric position, not name.  Useful for ordered XML like CallXML.
    var $trim_whitespace;   // variable to declare whether or not to keep whitespace inside xml tags

    // Read in XML on object creation.
    // We can take raw XML data, a stream, a filename, or a url.

    function XMLParser($data_source, $data_source_type='raw', $collapse_dups=0, $index_numeric=0, $trim_whitespace=0)
    {
        $this->collapse_dups   = $collapse_dups;
        $this->index_numeric   = $index_numeric;
        $this->trim_whitespace = $trim_whitespace;
        $this->data = '';

        // raw input
        if ($data_source_type == 'raw')
        {
            $this->data = $data_source;
        }
        elseif ($data_source_type == 'stream') {
            while (!feof($data_source))
            {
                $this->data .= fread($data_source, 1000);
            }
            // try filename, then if that fails...
        }
        elseif (file_exists($data_source))
        {
            $this->data = implode('', file($data_source));
        }
        // try url
        else
        {
            $fp = fopen($data_source,'r');
            while (!feof($fp))
            {
                $this->data .= fread($fp, 1000);
            }
            fclose($fp);
        }
    }

    // Parse the XML file into a verbose, flat array struct.
    // Then, coerce that into a simple nested array.
    function getTree()
    {
        $parser = xml_parser_create('UTF-8');
        xml_parser_set_option($parser, XML_OPTION_SKIP_WHITE, $this->trim_whitespace);
        xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 0);
        xml_parse_into_struct($parser, $this->data, $vals, $index);
        xml_parser_free($parser);

        $i = -1;
        return $this->getchildren($vals, $i);
    }

    // internal function: build a node of the tree
    function buildtag($thisvals, $vals, &$i, $type)
    {
        $tag=array();
        if (isset($thisvals['attributes']))
        {
            //$tag['attributes'] = $thisvals['attributes'];
            $tag = $thisvals['attributes'];
        }
        // complete tag, just return it for storage in array
        if ($type === 'complete')
        {
            // check if we're entering an empty value or not
            if (!empty($thisvals['value']))
            {
                // not empty, so lets assign $tag to the value element of $thisvals to avoid unnecessary obfuscation
                // doesn't work on cdata for some reason, oh well not a big deal
                $tag = $thisvals['value'];
            }
        }
        // open tag, recurse
        else
        {
            $tag = array_merge($tag, $this->getchildren($vals, $i));
        }
        return $tag;
    }

    // internal function: build an nested array representing children
    function getchildren($vals, &$i)
    {
        $children = array();     // Contains node data

        // Node has CDATA before it's children
        if ($i > -1 && !empty($vals[$i]['value']))
        {
            $children['value'] = $vals[$i]['value'];
        }

        // Loop through children, until hit close tag or run out of tags
        while (++$i < count($vals))
        {
            $type = $vals[$i]['type'];

            // 'cdata':     Node has CDATA after one of it's children
            //              (Add to cdata found before in this case)
            if ($type === 'cdata')
            {
                if (!empty($vals[$i]['value']))
                {
                    $children['value'] .= $vals[$i]['value'];
                }
            }

            // 'complete':  At end of current branch
            // 'open':      Node has children, recurse
            elseif ($type === 'complete' || $type === 'open')
            {
                $tag = $this->buildtag($vals[$i], $vals, $i, $type);
                if ($this->index_numeric)
                {
                    $tag['tag'] = $vals[$i]['tag'];
                    $children[] = $tag;
                }
                else
                {
                    $children[$vals[$i]['tag']][] = $tag;
                }
            }

            // 'close:      End of node, return collected data
            //              Do not increment $i or nodes disappear!
            elseif ($type === 'close')
            {
                break;
            }
        }
        if ($this->collapse_dups)
        {
            foreach($children as $key => $value)
            {
                if (is_array($value) && (count($value) == 1) && (!empty($value)))
                {
                    $children[$key] = $value[0];
                }
            }
            return $children;
        }
    }
}
?>

edit: I modified a copy of this to use curl to pull data from a external location with cookie support if you want that instead