From In-Portal Developers Guide

Revision as of 18:26, 18 April 2009

Вспомогательные классы
Статьи в этой категории
Обработка XML документов

XML is a convenient tree-structure format of organizing data. Each element can contain a value and/or other elements. Text inside an element can alternate with child elements, but it will still be considered a single value. Moreover, each element can have attributes. All of this is shown in the below example (let it be a file called "sample.xml"):

<?xml version="1.0" encoding="utf-8"?>
<our_document>
	<some_tag>
		Some tag's content.
	</some_tag>
	<planet_earth>
	    <continent id="1">
		North America
            	<country id="1">
			Canada
		</country>
          	<country id="2">
			USA
		</country>
	    </continent>
	    <continent id="2">
	    	Europe
            	<country id="1">
			Estonia
		</country>
            	<country id="2">
			Latvia
		</country>
            	<country id="3">
			Lithuania
		</country>
	    </continent>
	</planet_earth>
</our_document>

The CDATA Block

If the content of an element needs to contain symbols that could disrupt the correct parsing of an XML document, then the contents need to be inside a CDATA block. Some of these symbols include ">", "<" и "&". The syntax of a CDATA block looks like this:

<![CDATA[Some symbolic data, that > breaks & xml]]>

Data inside a CDATA block should not contain "]]>" because it will considered the end of the block. If it's absolutely necessary to have "]]>" then the following approach will work:

<![CDATA[]]]]><![CDATA[>]]>

In the above example, the contents are separated into two parts, each of which is inside a CDATA block. It's especially convenient to use CDATA in the situation where XML needs to be written to show to a user, but won't be processed when parsing the XML document.

Getting an XML Document

Before parsing an XML document, it has to be received.

Getting a Local File

To get a local file (i.e. located on the same computer as the site), the standard file_get_contents function can be used. This function returns the contents of file, the path of which is passed as its first argument.

$file_contents = file_get_contents(WRITEABLE . '/user_files/sample.xml');

Getting a Remote File

The file_get_contents function also allows retrieving the contents of a remote file. However, for security reasons, the server may restrict this functionality. Therefore, it's strongly recommended to use the standard K4 kCurlHelper class to retrieve remote files.

$curl_helper =& $this->Application->recallObject('CurlHelper');
/* @var $curl_helper kCurlHelper */
 
$xml_data = $curl_helper->Send('http://sample-host.com/sample.xml');

The Send method has a second optional parameter, which indicates that the connection should be terminated right after getting the contents of the document. By default, it's set to "true", i.e. terminate connection.

Parsing an XML Document

Realized in the kXMLHelper class is a convenient mechanism for parsing an XML document. The first step is to create an object of the class.

$xml_helper =& $this->Application->recallObject('kXMLHelper');
/* @var $xml_helper kXMLHelper */

Then, to start the parsing of the XML document retrieved earlier.

$root_node =& $xml_helper->Parse($file_contents);

A tree structure of objects will be returned as a result, in which all objects are connected using links. The contents of each object of the kXMLNode class looks like this:

kxmlnode Object (
	[Name] => xml_element_name
	[Attributes] => Array
	(
		[1st_attribute_name] => 1st_attribute_value
		[2nd_attribute_name] => 2nd_attribute_value
		...
	)
	[Children] => Array
	(
		[0] =>	kxmlnode Object
		[1] =>	kxmlnode Object
		...
	)
	[Data] => text_value_that_this_XML_element_encapsulates
	[firstChild] => kxmlnode Object
	[lastChild] => kxmlnode Object
	[Parent] =>  kxmlnode Object
	[Position] => 1
)

The below description of the contents of an object of the kXMLNode class will be based on the example at the beginning of the article.

The parent (root) object will be assigned to the $root_node variable, i.e. the xml-element object "our_document". The "Children" (private) attribute contains an array of all child elements of the current element. In this example, these are the two elements "some_tag" and "planet_earth". The last one has two "continent" child elements. It's import to understand that array elements are the same as other objects, they can themselves have child elements, etc.

Each element has a Position attribute. This is its order in the elements (same level elements, for example - "country" Canada and "country" USA). The "firstChild" and "lastChild" elements contain the first and last (from the point of view of their Position) child elements, respectively.

For further processing of received information the kXMLNode class methods and attributes are used.

Практическое использование kXMLHelper

Ниже приведён код, который распечатает все страны описанного выше XML-документа.

$root_node =& $xml_helper->Parse($xml_data);
/* @var $root_node kXMLNode */
 
// Getting first continent node
$continent_node =& $root_node->FindChild('continent');
 
// Cycling through it and all the rest continent nodes
do {
	// Getting first country node
	$country_node =& $continent_node->firstChild;
 
	// Cycling through it and all the rest continent nodes
    	do {
        	echo $continent_node->Attributes['ID'] . ' - ' . trim($continent_node->Data) . ': ' . $country_node->Attributes['ID'] . ' - ' . trim($country_node->Data) . '<br/>';
    	} while ($country_node =& $country_node->NextSibling());
} while ($continent_node =& $continent_node->NextSibling());

В переменной $continent_node сохраняется первый найденный объект "continent", т.е. - "North America". В первом цикле перебираются континенты. Для перехода к элементу того же уровня используется метод NextSibling (есть противоположный ему метод PrevSibling). Из континента выбирается первая страна. Внутренний цикл перебирает все страны данного континента и делает вывод в приведённом ниже формате. Как можно заметить, текстовое значение элементов доступно через атрибут "Data".

1 - North America: 1 - Canada
1 - North America: 2 - USA
2 - Europe: 1 - Estonia
2 - Europe: 2 - Latvia
2 - Europe: 3 - Lithuania

Методы класса "kXMLNode"

Ниже приведены public методы класса kXMLNode. Все методы рассчитаны на чтение данных, но не на их запись.