SELECT FLOOR(7 + (RAND() * 5));

However, when considering overhead, this is a rather expensive method to employ. This becomes glaringly apparent when dealing with large datasets.

One effective solution is to select the first and last ids, and let PHP generate the random integer (id).  Then, simply select the record using the random integer.

public function getRandomRecords()
{
	// Select the first id in the target table
	$statement = $this->db->prepare("SELECT some_id
	  FROM table_name
	  ORDER BY some_id ASC LIMIT 1");
	$statement->execute();
	$lowest_id = $statement->fetch(PDO::FETCH_ASSOC);

	// Select the last id in the target table
	$statement = $this->db->prepare("SELECT some_id
	  FROM table_name
	  ORDER BY some_id DESC LIMIT 1");
	$statement->execute();
	$highest_id = $statement->fetch(PDO::FETCH_ASSOC);

	$records_array = array();

	while(true)
	{
	  // Generate a random integer
	  $random_id = rand( $lowest_id['some_id'], $highest_id['some_id'] );

	  // Check to see if the record exists
	  $statement = $this->db->prepare("SELECT col_one, col_two, etc...
		FROM table_name
		WHERE some_id = {$random_id}";
		$statement->execute();
		$result = $statement->fetchAll(PDO::FETCH_ASSOC);

	  // If it exists, add it to the array
	  if($result)
	  {
	    $records_array[] = $record;
	  }

	  $i++;

	  // If the array contains 5 records, stop
	  if(count($records_array) == 5)
	  {
	    break;
	  }
	}

	return $records_array;
}

First Mistake

For my testing, I used a small subset of the data… weighing in at a measly 24 records.

Initially, all of my tests ran quite nicely. Then I decided to throw the complete (1 GB) XML file at it. Epic fail… I mean, it ran well for a while, but eventually ran out of memory. (And, yes… I did increase the memory_limit* to 1.5 GB and max_execution_time* to 5 hours.) I feared this may happen.

The problem with utilizing DOMDocument on large XML files is that it loads the data into an array. While parsing, that array is growing. Not good when you’re dealing with massive XML files.

With this fail under my belt, I went back to the drawing board. Knowledge is power… knowledge is power… knowledge is power.

My Next Move

XMLReader. From the PHP website: ”The XMLReader extension is an XML Pull parser. The reader acts as a cursor going forward on the document stream and stopping at each node on the way.” OK, sounds considerably more promising.

And Survey Says, Ding!

$file = "PATH_TO_FILE";

$reader = new XMLReader();

$reader->open($file);

while( $reader->read() )

{

// Execute processing here

}

$reader->close();

After that, it was gravy. Well, aside from the additional logic that had to go into it. That’s easily a topic all of it’s own, perfect for perhaps a “Part 2″ of this post. No promises though… unless, of course, incoming requests prompt for more information!

* How to modify PHP’s “memory_limit” and “max_execution_time” on a per script basis

// Tweak some PHP configurations
ini_set('memory_limit','1536M'); // 1.5 GB
ini_set('max_execution_time', 18000); // 5 hours

Microsoft Word. A necessary evil... shame, shame.

About a month ago, we ran into a problem with hyperlinks in a Microsoft Word document. Wait! What? H-y-p-e-r-l-i-n-k-s? Yup. The most basic operation known in the history of the internet; the fundamental building block for information sharing in today’s technological landscape.

So, we decided to build a function in a web application which outputs a Microsoft Word document complete with links. Here’s the kicker. When we finally succeeded in generating the Word document, we opened it and tried the links. This produced some weird results and basically failed when linking back to our web application.

OK, here’s the deal. Have you ever noticed that whenever you click on a link in a Microsoft Word document, a status bar appears at the bottom of the document window? Have you ever wondered what in the world it’s doing? I have, but never pursued finding out.

As it turns out, Microsoft Word first sends the request to one of Microsoft’s servers. Then, Microsoft’s server sends the request back to the intended location. The problem is that it sends the request back, which is now originating from Microsoft’s server, void of any session data. So, if the end-user is logged in with established session data, good luck in successfully routing the user to the target… at this point, the request is coming from another user-agent.

So, here’s what we had to do in PHP to remedy this rather perplexing situation:

if(stristr($_SERVER['HTTP_USER_AGENT'], 'Word'))
{
die();
}

One can only speculate what Microsoft is doing with all of these requests. For some reason, I’m leaning towards the word “harvest”. I surely hope this helps and saves time for anyone who runs into this issue! Microsoft #fail.

Rick's Fish & Pet Supply Website

Actually, this was a team effort. The visionary was my wife, Julia. I just helped with some of the heavier technical aspects. There’s more to be added to the website. This is just Phase 1. Visit Rick’s Fish and Pet Supply’s Website

MODx: Content Management System, Framework, and More

Don’t get me wrong, I love WordPress. However, I will say that the PHP programming behind WordPress is, how should I say this? Um… a little weak. So, I’m trying out another CMS, MODx. We’ll see… more on this as things develop.

UPDATE: Vanessa Robison’s website is now live! I must say, it looks great! Check it out!