ATWD1 Assignment: Code Example (1)

The original input file is a simple CSV file with 7 fields and 316 records. The first record is a header row.

We can process this CSV file and generate a simple XML file very easily.

	<?php
	$i = 0;
	$r = 0;
	$columns = 0;
	$file = "Quality_of_Life_-_Travel_to_Work_2013.csv";
	$xml = '';
	if (($handle = fopen($file, "r")) !== FALSE) {
	$xml .= '<transport>';
	while ($data = fgetcsv($handle, 2048, ",")) {
	$i++;
	if ($i == 1) {
	$columns = count($data);
	$header = $data;
	} else {
	$r++;
	$xml .= '<record id="' . $r . '">';
	# iterate over columns
	for ($c = 0; $c < $columns; $c++) {
	$xml .= "<$header[$c]>" . $data[$c] . "</$header[$c]>";
	}
	$xml .= '</record>';
	}
	}
	$xml .= '</transport>';
	}
	fclose($handle);
	# save the file as a DOM document
	$doc = new DOMDocument();
	$doc->loadXML($xml);
	$doc->save("transport.xml");
	# print out xml
	echo header("Content-type: text/xml");
	echo $xml;
	?>

view raw csv_to_xml.php hosted with ❤ by GitHub

Run

The output file isn't optimal for processing and also contains some repeating (unecessary) data. The location node repeats data already in the lat./long. nodes and the year value is repeated too.

However, we've taken a big leap. The data is now in a DOM compatible format!

For readabilty & easier processing purposes we would prefer a structure like the following:

	<?xml version="1.0" encoding="UTF-8"?>
	<transport year='2013'>
	<ward lat='51.46619' long='-2.583456' name='Ashley'>
	<mode type='car(as driver)' percent='25.3' />
	<mode type='car(as passenger)' percent='3.7' />
	<mode type='another' percent='0' />
	<mode type='bus' percent='10.9' />
	<mode type='car' percent='29' />
	<mode type='cycle' percent='20.4' />
	<mode type='moped/motorbike' percent='0.6' />
	<mode type='train' percent='1.4' />
	<mode type='walking' percent='37.7' />
	</ward>
	<!-- more wards -->
	</transport>

view raw transport_fragment.xml hosted with ❤ by GitHub

For this transformation we can use PHP's built-in DOM facade library simpleXML or use XSLT (built for such xml-to-xml transformations).

Unfortunatly current PHP builds don't provide native support for XPath 2.0/3.0 so we have to do with XPath 1.0.

XPath 2.0/3.0 provide very powerful built in functions like distinct-values() so an expression like distinct-values(//record/Ward) would generate an array with unique wards.

(We can get around this by installing and using an external library like Saxon but lets use what we have - the current UWE/CEMS setup and XPath 1.0.)

Let's first consider a PHP only solution using the simpleXML and XMLWriter modules & XPath 1.0.

	<?php
	# important
	@date_default_timezone_set("GMT");
	# create a simpleXML object with file as input
	$xml = simplexml_load_file("transport.xml");
	# generate a array holding unique wards
	$wards = $xml->xpath("//Ward[not(preceding-sibling::Ward)]");
	$ward = array_unique($wards);
	# create and intialise new XMLWriter object
	$writer = new XMLWriter();
	$writer->openURI('transport_v1.xml');
	$writer->startDocument("1.0");
	$writer->setIndent(4);
	# get the year value
	$year = (string) $xml->xpath("//Year")[0];
	# start writing the new xml
	$writer->startElement('transport');
	$writer->writeAttribute('year', $year);
	# walk the $ward array & process data
	foreach ($ward as $w) {
	# get an array of records matching current ward
	$records = $xml->xpath("//*[Ward='$w']");
	# get the geo-code
	$lat = (string) $records[0]->Approx_lat;
	$long = (string) $records[0]->Approx_long;
	# start the ward element
	$writer->startElement('ward');
	$writer->writeAttribute('lat', $lat);
	$writer->writeAttribute('long', $long);
	$writer->writeAttribute('name', $w);
	# walk the records array & process data
	foreach ($records as $record) {
	# start the mode element
	$writer->startElement('mode');
	switch (true) {
	case strpos((string) $record->Question, "as driver"):
	$writer->writeAttribute('type', 'car (as driver)');
	$writer->writeAttribute('percent', $record->Value);
	break;
	case strpos((string) $record->Question, "as passenger"):
	$writer->writeAttribute('type', 'car (as passenger)');
	$writer->writeAttribute('percent', $record->Value);
	break;
	case strpos((string) $record->Question, "another"):
	$writer->writeAttribute('type', 'another');
	$writer->writeAttribute('percent', $record->Value);
	break;
	case strpos((string) $record->Question, "bus"):
	$writer->writeAttribute('type', 'bus');
	$writer->writeAttribute('percent', $record->Value);
	break;
	case strpos((string) $record->Question, "work by car"):
	$writer->writeAttribute('type', 'car');
	$writer->writeAttribute('percent', $record->Value);
	break;
	case strpos((string) $record->Question, "cycle"):
	$writer->writeAttribute('type', 'cycle');
	$writer->writeAttribute('percent', $record->Value);
	break;
	case strpos((string) $record->Question, "moped/motorbike"):
	$writer->writeAttribute('type', 'moped/motorbike');
	$writer->writeAttribute('percent', $record->Value);
	break;
	case strpos((string) $record->Question, "train"):
	$writer->writeAttribute('type', 'train');
	$writer->writeAttribute('percent', $record->Value);
	break;
	case strpos((string) $record->Question, "walking"):
	$writer->writeAttribute('type', 'walking');
	$writer->writeAttribute('percent', $record->Value);
	break;
	}
	$writer->endElement();
	}
	$writer->endElement();
	}
	# end the document and flush to file
	$writer->endDocument();
	$writer->flush();
	?>

view raw tdom_to_xml.php hosted with ❤ by GitHub

You can view the generated file transport_v1.xml

We'll next use XSLT 1.0 to do the same task and take a look at XSLT 2.0/3.0.

---

url: http://www.cems.uwe.ac.uk/~p-chatterjee/2015-16/modules/atwd1/assignment/examples/assignment_code_example1.html