|
University of Oxford |
Computing Services |
|
Processing XML using PHP |
|
| Author: Barry Cornelius Date: 8th November 2002; first created 4th November 2002 |
<table border=1>
<tr>
<td><b>inkjet cartridges</b></td>
<td><em>HP Deskjet print cartridge 51645A</em></td>
<td bgcolor="yellow">19.50</td>
</tr>
...
</table>
<consumables>
<product>
<category>inkjet cartridges</category>
<item>HP Deskjet print cartridge 51645A</item>
<price>19.50</price>
</product>
...
</consumables>
http://www.dur.ac.uk/barry.cornelius/papers/xml.processing.using.php/code/consumables.xml.
<price>19.50</price>is called an element. It is introduced by a start tag. In this example, the start tag is:
<price>And it is terminated by an end tag:
</price>A start tag may have one or more attributes as in:
<price units="pounds" country="UK">19.50</price>An element may be empty as in:
<price amount="19.50"></price>and this can be abbreviated to:
<price amount="19.50"/>Another example is:
<unknown_price></unknown_price>and this can be abbreviated to:
<unknown_price/>
As I see it, there are three main goals of XML:
<!ELEMENT consumables (product*)> <!ELEMENT product (category, item, price)> <!ELEMENT category (#PCDATA)> <!ELEMENT item (#PCDATA)> <!ELEMENT price (#PCDATA)>http://www.dur.ac.uk/barry.cornelius/papers/xml.processing.using.php/code/consumables.dtd.
More recently, XML schemas have been introduced. They can be used to define the structure of some XML (instead of using a DTD). Two of the main differences are:
<xsl:template match="consumables">
<html>
<body>
<table>
<xsl:apply-templates/>
</table>
</body>
</html>
</xsl:template>
<xsl:template match="product"> <tr> <td><xsl:value-of select="category"/></td> <td><xsl:value-of select="item"/></td> <td><xsl:value-of select="price"/></td> </tr> </xsl:template>http://www.dur.ac.uk/barry.cornelius/papers/xml.processing.using.php/code/consumables.xsl.
Three approaches for processing an XML document will now be considered.
We could write a program to process an XML document. The program could be used to produce HTML which we could then make available on the WWW. More details about this are given in Section 5.
The problem with using a tool is that we have to remember to run it everytime the XML document is updated. The remaining two approaches enable the user of a browser to work directly on an XML document.
If a webserver supports the running of server-side programs, then a browser visiting a page could trigger a webserver to run some code that reads an XML document and produces HTML. This could be done by using a CGI program or by running a servlet (coded in Java) (if your webserver supports the running of servlets). However, Section 6 of this document looks at how a PHP script can be used to read an XML document and produce HTML.
From the early days of XML, it has been reasonably easy to write programs that read XML documents. To begin with, these programs were mainly written in Java, but now it is easy to write such programs in a number of programming languages (including PHP).
Although the code of a PHP file is normally triggered into action by a webserver, it is now possible to run standalone PHP programs, i.e., to run PHP programs from a Unix prompt or from a Windows command shell window. Although the support for this is experimental in PHP 4.2.3, it is fully supported in PHP 4.3.0 (which is currently only in pre-release). We now look at how to write a PHP program that processes an XML document.
The XML parser that comes with PHP allows you to write PHP code that walks through an XML document. It generates events for each part of a document, generating an event for each start tag, for each end tag, and for each occurence of characters appearing between a start tag and an end tag. In a PHP program, you can nominate functions to be called when these events occur.
Let's look at this in more detail. In a PHP program, you first need to create an XML parser:
$xmlparser = xml_parser_create();You then register the functions that you wish to be called. For example, given:
xml_set_element_handler($xml_parser, "startElement", "endElement"); xml_set_character_data_handler($xml_parser, "characterData");the startElement function will be called when the parser meets a start tag; endElement will be called when it meets an end tag; and characterData will be called for the characters in between.
Having done all that, you then set the XML parser loose on your document:
$fp = fopen("consumables.xml", "r");
while ($data = fread($fp, 4096)) {
xml_parse($xml_parser, $data, feof($fp));
}
Here is a complete program that reads consumables.xml and outputs an HTML table:
<%
function startElement($parser, $name, $attrs) {
if ($name == "consumables") {
print "<table>";
}
else if ($name == "product") {
print "<tr>";
}
else {
print "<td>";
}
}
function endElement($parser, $name) {
if ($name == "consumables") {
print "</table>";
}
else if ($name == "product") {
print "</tr>";
}
else {
print "</td>";
}
}
function characterData($parser, $data) {
print $data;
}
function startDocument() {
print "<html>\n";
print "<body>\n";
}
function endDocument() {
print "</body>\n";
print "</html>\n";
}
$file = "consumables.xml";
$xml_parser = xml_parser_create();
// insist that the tags have the right case
xml_parser_set_option($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");
}
startDocument();
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);
endDocument();
%>
The xml_parse function walks through an XML document calling your three functions as it reaches appropriate points of the XML document.
Note that each part of the XML document is visited only once. If instead you want the ability to wander around an XML document, it is better to use the xml_parse_into_struct function: this will put a representation of the XML document into an array. PHP also has some DOM XML functions, but these are currently experimental.
Instead of sprinkling the HTML we want to generate all over a PHP program, we could instead use the XSL stylesheet that was produced in Section 3.
PHP provides an extension that can do XSL transformations. In order to use this extension, the person installing PHP needs to build it with a library that implements XSLT. Currently, the only library supported by PHP is the Sablotron library, but future versions of PHP will support other libraries (such as Xalan and libxslt).
No matter what library for XSLT has been included, it will be used in the same way. Its use is demonstrated by the following program.
<%
$xsltproc = xslt_create();
$html = xslt_process($xsltproc,"consumables.xml","consumables.xsl");
if (!$html) {
die("XSLT processing error:" . xslt_error($xsltproc));
}
xslt_free($xsltproc);
echo $html;
%>
The crucial statement of this program is:
$html = xslt_process($xsltproc,"consumables.xml","consumables.xsl");The call of the xslt_process function takes the file consumables.xml and transforms it using the rules in the file consumables.xsl. This will produce a long string of HTML instructions, and this is stored in the $html variable. In the above program, the statement:
echo $html;causes this HTML to be output.
Instead of running a tool to transform an XML document into HTML, you could get a webserver to run a program that does the transformation. This can be done in many ways, including the use of a CGI script, a PHP script, a servlet using SAX, or by using a JavaServer page.
For example, we could have the following file of PHP code.
<%
$xsltproc = xslt_create();
$html = xslt_process($xsltproc,"consumables.xml","consumables.xsl");
if (!$html) {
die("XSLT processing error:" . xslt_error($xsltproc));
}
xslt_free($xsltproc);
echo $html;
%>
If this file is put into a public_html directory, we are providing a PHP script that automatically generates HTML from an XML document.
http://webster.dur.ac.uk/barry.cornelius/myphp/xslt/small.php.Accessing a file of XML assumes that some other process has already created it (and keeps it up to date). It may be more useful for this XML to be generated by some other means. If we want to do this, we have to call the xslt_process function in a different way. Just to demonstrate how this is done, here is a PHP script that has some XML stored in a string variable:
<%
$xml =
"<?xml version=\"1.0\"?>"
. "<consumables>"
. " <product>"
. " <category>cassette tape</category>"
. " <item>DC6150 cartridge</item>"
. " <price>9.50</price>"
. " </product>"
. "</consumables>";
$arguments = array("/_xml" => $xml);
$xsltproc = xslt_create();
$html = xslt_process($xsltproc, "arg:/_xml", "consumables.xsl", NULL, $arguments);
if (!$html) {
die("XSLT processing error:" . xslt_error($xsltproc));
}
xslt_free($xsltproc);
echo $html;
%>
http://webster.dur.ac.uk/barry.cornelius/myphp/xslt/embeddedxml.php.
It may be that the data we want is stored in a database. What we could do is as follows:
For example, suppose we want to offer a WWW page where the visitor can look at a database containing some items displaying those items that have a cost less than some price. We will provide a WWW form for the visitor to type in the price.
Suppose we also want to offer the visitor a choice on how to display the data. We could provide radio buttons on the WWW form to get the visitor to choose between different styles of layout.
http://webster.dur.ac.uk/barry.cornelius/myphp/xslt/prices.htm.
Here is a WWW form that achieves this:
<HTML>
<BODY>
<FORM METHOD="POST" ACTION="prices.php">
What is your choice of price?<BR>
<INPUT TYPE="text" NAME="somevalue">
<BR>
<INPUT TYPE="radio" NAME="layout" VALUE="plain"> plain
<BR>
<INPUT TYPE="radio" NAME="layout" VALUE="table"> table
<BR>
<INPUT TYPE="submit" VALUE="Submit price">
<INPUT TYPE="reset">
</FORM>
</BODY>
</HTML>
When the visitor to this form clicks on the Submit price button, the prices.php script will be executed with two variables already set: the $somevalue variable will be initialised to the value that was typed in the form's textbox, and the $layout variable will be initialised to either plain or table.
We can use the value of the $somevalue variable in an SQL query (that produces a resultset):
mysql_connect("mysql.dur.ac.uk", "", "");
$SQLQuery = "SELECT * FROM consum WHERE price < $somevalue";
$result = mysql_db_query("Pdcl0bjc_prices", $SQLQuery);
$numrows = mysql_numrows($result);
for ($rownum = 0; $rownum<$numrows; $rownum++) {
echo mysql_result($result,$rownum,"ID");
echo mysql_result($result,$rownum,"goods");
echo mysql_result($result,$rownum,"price");
echo "<BR>\n";
}
Each iteration of the above for loop outputs three values. Instead of using the for loop to output these values, we could get it to produce XML-decorated forms of these values, e.g. to generate something like:
<product> <ID>42</ID> <goods>3.5" floppy disc</goods> <price>0.35</price> </product>
And we can get it to append these characters to the end of a string (being stored in a variable called $xml). This is illustrated by the prices.php script:
<%
mysql_connect("mysql.dur.ac.uk", "", "");
$SQLQuery = "SELECT * FROM consum WHERE price < $somevalue";
$result = mysql_db_query("Pdcl0bjc_prices", $SQLQuery);
$numrows = mysql_numrows($result);
if ( $numrows == 0) {
die("<p>There are no consumables with a price < $somevalue</p>");
}
$xml = "<?xml version=\"1.0\"?><consumables>";
for ($rownum = 0; $rownum<$numrows; $rownum++) {
$xml .= "<product>";
$xml .= "<ID>".mysql_result($result,$rownum,"ID")."</ID>";
$xml .= "<goods>".mysql_result($result,$rownum,"goods")."</goods>";
$xml .= "<price>".mysql_result($result,$rownum,"price")."</price>";
$xml .= "</product>";
}
$xml .= "</consumables>";
$arguments = array("/_xml" => $xml);
$xsltproc = xslt_create();
$html = xslt_process($xsltproc, "arg:/_xml", $layout.".xsl", NULL, $arguments);
if (!$html) {
die("XSLT processing error:" . xslt_error($xsltproc));
}
xslt_free($xsltproc);
echo $html;
%>
The above script arranges for the value of
$xml
to be passed to a call of
xslt_process.
In this call,
the third argument
is
either
the value
"plain.xsl"
or
the value
"table.xsl".
The file plain.xsl could contain:
<?xml version="1.0"?>
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output method="html"/>
<xsl:template match="consumables">
<html>
<body>
<pre>
<xsl:apply-templates/>
</pre>
</body>
</html>
</xsl:template>
<xsl:template match="product">
<xsl:value-of select="ID"/>
<xsl:value-of select="goods"/>
<xsl:value-of select="price"/>
<br/>
</xsl:template>
</xsl:stylesheet>
and the file table.xsl could contain:
<?xml version="1.0"?>
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output method="html"/>
<xsl:template match="consumables">
<html>
<body>
<table>
<tr bgcolor="yellow">
<td>ID</td>
<td>goods</td>
<td>price</td>
</tr>
<xsl:apply-templates/>
</table>
</body>
</html>
</xsl:template>
<xsl:template match="product">
<tr>
<td><xsl:value-of select="ID"/></td>
<td><xsl:value-of select="goods"/></td>
<td><xsl:value-of select="price"/></td>
</tr>
</xsl:template>
</xsl:stylesheet>
As time moves on, WWW browsers play catch-up with new developments. We now look at the support for XML and XSL that is provided in recent versions of browsers: we look at version 6 of Microsoft's Internet Explorer (IE6) and version 7 of Netscape's Navigator (NS7).
Both IE6 and NS7 provide support for XML with CSS1 (version 1 of Cascading Style Sheet). To use CSS, you need to augment an XML document with a xml-stylesheet processing instruction. This results in the text that is in the file http://www.dur.ac.uk/barry.cornelius/papers/xml.processing.using.php/code/css1.xml.
<?xml version="1.0" standalone="no"?> <?xml-stylesheet type="text/css" href="css1.css"?> <!DOCTYPE consumables SYSTEM "consumables.dtd"> <consumables> ... </consumables>The type attribute of this xml-stylesheet processing instruction says that this XML document is to be transformed using a CSS stylesheet and the href attribute gives the URL of the file containing the actual CSS instructions.
The file css1.css might contain:
consumables { display: block }
product { display: block }
category { font-size: x-large;
color: black }
item { font-size: large;
color: red }
price { background-color: yellow;
color: blue }
CSS2 introduces some new aspects like the handling of tables. So instead we could use the file http://www.dur.ac.uk/barry.cornelius/papers/xml.processing.using.php/code/css2.xml.
<?xml version="1.0" standalone="no"?> <?xml-stylesheet type="text/css" href="css2.css"?> <!DOCTYPE consumables SYSTEM "consumables.dtd"> <consumables> ... </consumables>where the file css2.css contains:
consumables { display: table }
product { display: table-row }
category { display: table-cell; font-size: x-large;
color: black }
item { display: table-cell; font-size: large;
color: red; text-indent: 0.1in }
price { display: table-cell; background-color: yellow;
color: blue; text-align: right; text-indent: 0.1in }
Although these instructions are obeyed correctly by
NS7,
they are not understood by
IE6.
Both IE6 and NS7 can use XSL to process an XML document. You need to insert the following xml-stylesheet processing instruction in your XML document:
<?xml-stylesheet type="text/xsl" href="consumables.xsl"?>
The files http://www.dur.ac.uk/barry.cornelius/papers/xml.processing.using.php/code/xsl.xml and http://www.dur.ac.uk/barry.cornelius/papers/xml.processing.using.php/code/consumables.xsl contain the texts that can be used by both NS7 and IE6.