Sunday, September 18, 2011

Top-10 PHP String Functions

By day I work as a programmer at ShoreGroup, Inc. By night I'm a freelance developer and now the managing editor for SitePoint's latest site, PHPMaster.com. Helping out with the site has been pretty fun so far; my Australian counterparts are all pretty cool, and I've met some really great new authors too. If you haven't visited yet, take a moment and check out PHPMaster.com (there's still some wrinkles to iron out on the site, but we're working to identify and fix them all as soon as we can).

Part of my duties as a managing editor include working with authors to make sure the site's content is well balanced. PHPMaster.com is targeting PHP programmers of all skill levels, so there should be a good mix of basic, beginner, intermediate, and advanced content. Planning for a beginner article that demonstrates basic string handling functions, I wondered which function to highlight. I wanted to show ones that would be most relevant, not necessarily ones that were my favorite, so I decided to do some static analysis of popular open-source projects to find out which string functions were used the most. The results were surprising, so I thought I'd share my "research."

I used the source of a closed-source PHP project that I have access to and the following open-source (or open-source-ish) projects as code samples for the analysis:

Then I ran the following PHP to tally the functions:

#! /usr/bin/env php
<?php
if ($_SERVER["argc"] != 4) {
    $script = basename(__FILE__);
    fprintf(STDERR, "usage: %s directory max exts\n", $script);
    fprintf(STDERR, "\tdirectory - directory to start traversal\n");
    fprintf(STDERR, "\tmax - maximum number of results to return\n");
    fprintf(STDERR, "\texts - comma-separated list of file extensions\n");
    fprintf(STDERR, "example: %s /var/www 20 php,inc\n", $script);
    exit(1);
}
// no error-checking... don't be stupid
$directory = $_SERVER["argv"][1];
$max = $_SERVER["argv"][2];
$extsRegex = "/(" . str_replace(",", "|", $_SERVER["argv"][3]) . ')$/';

$dirIter = new RecursiveDirectoryIterator($directory);
$recIter = new RecursiveIteratorIterator($dirIter);
$iter = new RegexIterator($recIter, $extsRegex);

$funcs = array();
foreach ($iter as $file) {
    $tokens = token_get_all(file_get_contents($file));
    foreach ($tokens as $t) {
        if (is_array($t) && $t[0] == T_STRING && function_exists($t[1])) {
            if (!isset($funcs[$t[1]])) {
                $funcs[$t[1]] = 0;
            }
            $funcs[$t[1]]++;
        }
    }
}
arsort($funcs);

$max = min(count($funcs), $max);
if ($max) {
    list($funcs) = array_chunk($funcs, $max, true);
}
print_r($funcs);

I took the resulting list of functions and extracted the string-specific ones to come up with this top-10 list (sorted in decreasing order of most-used):

  1. substr() - 6,605
  2. sprintf() - 5,604
  3. implode()/join() - 4,829
  4. strlen() - 4,557
  5. chr() - 4,122
  6. str_replace() - 4,009
  7. explode() - 3,401
  8. strpos() - 3,238
  9. htmlspecialchars() - 3,171
  10. trim() - 2,998

I expected functions like substr() and trim() to be on the list, but chr() was a surprise. Before this I probably would have laughed at you if you told me chr() is used almost twice as much as strtolower() (which came in 12th place with 2,267). Interesting results indeed!