Question

I have a class that is building some HTML using data stored in an array. There are around 100 items in this array. Each item includes information like company name, a description, and flags for the different programming languages the company supports. I am doing string concatenation as I build the HTML for each item.

I have noticed that performance suddenly takes a huge hit when I append the programming language data. I see the page rendering timer jump from 0.15 secs to ~0.60 secs. This time includes grabbing the same data from the database each time. I can consistently get the performance to jump between these 2 times but commenting/uncommenting the following line of code:

$html .= '<div class="programmingLanguages"><strong>Programming Languages</strong> '.implode(', ', $progLanguagesArray).'</div>';

I’ve also been able to get the same performance drop by appending a long test string, something like this:

$html .= 'testtesttesttesttesttesttesttesttesttesttesttesttesttesttesttesttest';

What’s especially bizarre is that I have another line of code that uses the same ‘implode’ function and does NOT make any significant difference in performance:

$html .= '<div class="integrationMethods"><strong>Integration Methods:</strong> '.implode(', ', $intMethodsArray).'</div>';

Does anybody have any insight into what might be going on here? I am doing tons of concatenation like this elsewhere in my code and haven’t seen anything like this before. At this point, I’m stumped.

Here’s the full class:

class DeveloperView {

public static function getHtml($developers) {
    $html = '';
    $html .= '<div>';
    $html .= '<div>';

    $count = 0;
    foreach ($developers as $developer) {
        $url = $developer['attributes']['url'];
        $phone = $developer['attributes']['phone'];
        $company = $developer['attributes']['desc'];
        $active = $developer['attributes']['active'];
        $desc = $developer['object_value'];

        $intMethodsArray = array();
        if ($developer['attributes']['m1']) { $intMethodsArray[] = 'method 1'; }
        if ($developer['attributes']['m2']) { $intMethodsArray[] = 'method 2'; }
        if ($developer['attributes']['m3']) { $intMethodsArray[] = 'method 3'; }
        if ($developer['attributes']['m4']) { $intMethodsArray[] = 'method 4'; }
        if ($developer['attributes']['m5'])     { $intMethodsArray[] = 'method 5'; }

        $progLanguagesArray = array();
        if ($developer['attributes']['dotnet']) { $progLanguagesArray[] = '.Net (C# or VB.Net)'; }
        if ($developer['attributes']['asp'])    { $progLanguagesArray[] = 'Classic ASP'; }
        if ($developer['attributes']['cf'])     { $progLanguagesArray[] = 'Cold Fusion'; }
        if ($developer['attributes']['java'])   { $progLanguagesArray[] = 'Java'; }
        if ($developer['attributes']['php'])    { $progLanguagesArray[] = 'PHP'; }
        if ($developer['attributes']['perl'])   { $progLanguagesArray[] = 'Perl'; }
        if ($developer['attributes']['other'])  { $progLanguagesArray[] = 'Other'; }

        $html .= '<div class="';
        if ($count % 2 == 0) {
            $html .= 'listingalt';
        } else {
            $html .= 'listing';
        }
        $html .= '">';

        $html .= '<div class="developerPhone">'.$phone.'</div>';
        $html .= '<a class="ext_link" target="_blank" href="'.$url.'">'.$company.'</a>';

        $html .= '<div>';
        if (!empty($intMethodsArray)) {
            $html .= '<div class="integrationMethods"><strong>Integration Methods:</strong> '.implode(', ', $intMethodsArray).'</div>';
        }

        if (!empty($progLanguagesArray)) {
            $html .= '<div class="programmingLanguages"><strong>Programming Languages</strong> '.implode(', ', $progLanguagesArray).'</div>';
        }

        $html .= '</div>';

        $html .= '<p>'.$desc.'</p>';

        $html .= '</div>'."\n";

        $count++;
    }

    $html .= '</div></div>';

    return $html;
}

}

Answer 1

Now that I can provide an answer, I’ll just post my follow-up comment as the ‘answer’…

I did indeed have a ‘bug’ in my timer, in that it was calculating the end processing time AFTER the echo of the HTML. So the amount of data being sent to the browser was effecting the processing time, where I was expecting to see the time spent processing BEFORE transmitting any data.