Generating Reports and Statistics in PHP : Page 2

The XML_Statistics PEAR Package
To analyze XML documents, you can install and use the XML_Statistics package, which provides methods for obtaining statistics about tags, attributes, entities, processing instructions, data blocks, and CDATA sections for any well-formatted XML document.

To install the PEAR, use this command (version 0.1 is the beta version):

   > pear install -alldeps XML_Statistics-0.1

After installing the package, you'll find the code for the base class of the XML_Statistics package in the file XML_Statistics.php. Here are some of its main functions:

• boolean analyzeFile(mixed $file,string $filename): This function analyzes an XML file by loading it from a file path or URL. To see the results of the analysis use the countX() and getX() methods.
• integer countTag([string $tagname = null]): This function returns the number of occurrences of a tag in the entire XML document. The tag name is passed to the function through the $tagname argument.
• integer countAttribute([string $attribute = null], [string $tagname = null]): This function returns the number of occurrences of an attribute. You pass the attribute name to the function via the $attribute argument. If you don't specify the second argument, $tagname, then the function searches for the specified attribute in the entire XML document; otherwise, it limits the search range to the specified tag.
• integer getMaxDepth(): This function returns the maximum nesting level in the document, the "depth" of the XML tree.
• integer countTagsInDepth(integer $depth): This function returns the number of tags that "live" at the specified depth. The root tag depth is zero.
• integer countExternalEntity([string $name = null]): This function returns the number of occurrences of external entities. If you don't specify an entity name then the function counts the total number of external entities; otherwise, it counts only the occurrences of the specified entity.
• integer getCDataLength(): This function return the total length of all CDATA sections.

Listing 1 contains an XML document used for test purposes in the example code (see the file myxml.xml in the downloadable code).

The following PHP example uses the XML document in Listing 1 to retrieve some basic statistics about tags, attributes, and CDATA sections:

   <?php
     //import Statistics.php
     require_once 'XML/Statistics.php';
       
     //ignore whitespaces
     $stat = new XML_Statistics(array("ignoreWhitespace" => true));
       
     //analyze a file or URL
     $result = $stat->analyzeFile("myxml.xml");
       
     if ($stat->isError($result)) {
       die("Error: " . $result->getMessage());
     } 
     else 
     {        
       // total number of tags
       echo "Total tags: " . $stat->countTag()."<br>";
       
       // count number of 'type' attribute
       echo "Occurences of attribute type: " . $stat->countAttribute("type")."<br>";
             
       // get the maximum depth
       echo "Maximum depth: " . $stat->getMaxDepth()."<br>";
   
       // count total number of tags in depth 3
       echo "The number of tags in depth 3: " . $stat->countTagsInDepth(3)."<br>";
       
       // count the occurences of data blocks
       echo "Data chunks: " . $stat->countDataChunks()."<br>";
   
       // get the length of all CData sections
       echo "Length of all data chunks: " . $stat->getCDataLength()."<br>";
     }
   ?>

The output of this example is:

   Total tags: 16
   Occurences of attribute type: 2
   Maximum depth: 3
   The number of tags in depth 3: 10
   Data chunks: 6
   Length of all data chunks: 93

You can combine the XML statistics with statistical functions such as max, min, midrange, sum, variance, quartiles, etc. by installing the Math_stats PEAR package, which is usually used in conjunction with XML_Statistics PEAR. You install the Math_stats PEAR like this:

   pear install Math_stats

The most-commonly-used functions of the Math_stats package are:

• mixed calcBasic([boolean $returnErrorObject = true]): This function calculates a basic set of statistics.
• mixed calcFull([boolean $returnErrorObject = true]): This function calculates a full set of statistics.

The next example demonstrates both these functions using the myxml.xml document in Listing 1:

   
   <?php
   
     //import Statistics.php and Stats.php
     require_once 'XML/Statistics.php';
     require_once 'Math/Stats.php';
   
     $stat = new XML_Statistics();
     $result = $stat->analyzeFile("myxml.xml");
       
     if ($stat->isError($result)) {
       die("Error: " . $result->getMessage());
     } 
     else {
       // get the number of tags per tagname
       $tags = $stat->getTagOccurences();
   
       // use Math_Stats class
       $stats = new Math_Stats();
         
       // set the data
       $stats->setData($tags);
   
       // calculates a basic set of statistics
       $stats1 = $stats->calcBasic();
   
       // calculates a full set of statistics
       $stats2 = $stats->calcFull();
   
       echo "<pre>";
       print_r('<b><u>'."A basic set of statistics".'</u></b><br /><br />');
   
       // print a basic set of statistics 
       print_r($stats1);
   
       print_r('<br /><b><u>'."A full set of statistics".'</u></b><br /><br />');
   
       // print a full set of statistics
       print_r($stats2);
   
       echo "</pre>";
     }
   ?>
   

The "basic statistics" portion of the output of this example is shown below. Listing 2 shows the full output.

   A basic set of statistics
   Array
   (
       [min] => 1
       [max] => 2
       [sum] => 16
       [sum2] => 20
       [count] => 14
       [mean] => 1.1428571428571
       [stdev] => 0.36313651960128
       [variance] => 0.13186813186813
   )

Now that you've seen how to generate reports and statistics from flat files, you can move on to generating them from data stored in relational databases.