The SAX Model

SAX is referred to as an event-driven model or system. SAX parses an XML document and as it encounters the different parts of an element it performs different functions. The different functions performed depend on how the programmer sets up the code.

There are three main elements that are interacted with – the start tag, the end tag, and the data between the tags. For example, when a parser encounters a start tag, like <City>, the parser runs a function that renders the tag with any defined XML, like <td>. That is, SAX can parse XML documents to substitute HTML for XML tags. Also, an XML document can be analyzed while parsing it with SAX to produce a subset, before transforming it to XML or HTML.

This means that while parsing, an event occurs when SAX encounters a pre-defined part of an XML document. Each of these events is defined with a handler. There are a total of seven different kinds of events within SAX that PHP supports with defined handlers:

PHP Function to Set Handler	Event Description
xml_set_element_handler()	Used to define the events for working with or 'handling' the start and end tags within XML. Element events are issued whenever the XML parser encounters start or end tags. There are separate handlers for start tags and end tags.
xml_set_character_data_handler()	Used to define the event for 'handling character, for example, swapping HTML for a start tag. Character data is roughly all the non-markup contents of XML documents, including the whitespace between tags. Note that the XML parser does not add or remove any whitespace; it is up to the programmer to decide whether the whitespace is significant.
xml_set_processing_instruction_handler()	PHP programmers should be familiar with processing instructions (PIs) already. is a processing instruction, where PHP is called the 'PI target'. The handling of these is application-specific, except that all PI targets starting with "XML" are reserved.
xml_set_default_handler()	This handler is the default and should always be used. It will be called for each piece of XML that doesn't have a set handler. The structure is similar to the switch structure in PHP, where this is the default case.
xml_set_unparsed_entity_decl_handler()	This handler will be called when an unparsed (NDATA) entity is found in the XML.
xml_set_notation_decl_handler()	This handler is called when a notation is found in the XML document.
xml_set_external_entity_ref_handler()	This handler is called when the XML parser finds a reference to an external parsed general entity. This can be a reference to a file or URL. For a demonstration of the external entity example, refer to: http://www.php.net/manual/en/ ref.xml.php#example.xml-external-entity

The SAX API doesn't allow for writing XML. So, here we will work with the three handlers that can read an XML file:

• xml_set_element_handler()
• xml_set_character_data_handler()
• xml_set_default_handler()

Using PHP's SAX Support

Support for SAX is built into PHP by default in the form of the Expat extension. Expat allows programmers to parse XML, either from strings or files, and create XML parsers. Expat doesn't allow for validating or checking for well-formedness of XML.

If the XML file is not well-formed, SAX will process as much of the file as it can, up to the point of error. When it encounters an error, Expat will spit out error messages like this:

XML error: mismatched tag at line 4

The exact error message depends on how you set up error handling in your code. There are over twenty different error codes that can be returned.

SAX does not write XML. To write XML, two classes are available:

• xmlwriterclass (http://freshmeat.net/projects/xmlwriterclass/) by Manuel Lemos – writes well-formed valid XML documents to the browser
• XMLFile by Chris Monson – writes well-formed valid XML documents to a file

Alternatively you can write your own custom code to write an array to the file system formatted as XML.

Verifying XML Support
Professional PHP4 Programming
SAX Example Code Page 28