CODEWITHSUNDEEP
phpperformancesubstr
bcsteeve
bcsteeve

Reputation: 1001

most efficient means of parsing a simple clearly defined string?

I'm only asking because this is looping millions of times.

string is simply like this:

01-20

Its always like that... 2 digits (leading zero) followed by hyphen and another 2 digits (leading zero). I simply need to assign the first (as integer) to one variable and the second (as integer) to another variable.

str_split? substr? explode? regex?

Upvotes: 2

Views: 85

Answers (2)

trincot
trincot

Reputation: 351128

Given a variable $txt, this has the best performance:

    $a = (int)$txt;
    $b = (int)substr($txt, -2);

You could measure the performance of different alternatives with a script like this:

<?php
$txt = "01-02";
$test_count = 4000000;

// SUBSTR -2
$time_start = microtime(true);
for ($x = 0; $x <= $test_count; $x++) {
    $a = (int)$txt; // numeric conversion ignores second part of string.
    $b = (int)substr($txt, -2); 
}
$duration = round((microtime(true) - $time_start) * 1000);
echo "substr(s,-2): {$a} {$b}, process time: {$duration}ms <br />";

// SUBSTR 3, 2
$time_start = microtime(true);
for ($x = 0; $x <= $test_count; $x++) {
    $a = (int)$txt; // numeric conversion ignores second part of string.
    $b = (int)substr($txt, 3, 2);   
}
$duration = round((microtime(true) - $time_start) * 1000);
echo "substr(s,3,2): {$a} {$b}, process time: {$duration}ms <br />";

// STR_SPLIT
$time_start = microtime(true);
for ($x = 0; $x <= $test_count; $x++) {
    $arr = str_split($txt, 3);
    $a = (int)$arr[0]; // the ending hyphen does not break the numeric conversion
    $b = (int)$arr[1];
}
$duration = round((microtime(true) - $time_start) * 1000);
echo "str_split(s,3): {$a} {$b}, process time: {$duration}ms <br />";

// EXPLODE
$time_start = microtime(true);
for ($x = 0; $x <= $test_count; $x++) {
    $arr = explode('-', $txt);
    $a = (int)$arr[0];
    $b = (int)$arr[1];
}
$duration = round((microtime(true) - $time_start) * 1000);
echo "explode('-',s): {$a} {$b}, process time: {$duration}ms <br />";

// PREG_MATCH
$time_start = microtime(true);
for ($x = 0; $x <= $test_count; $x++) {
    preg_match('/(..).(..)/', $txt, $arr);
    $a = (int)$arr[1];
    $b = (int)$arr[2];
}
$duration = round((microtime(true) - $time_start) * 1000);
echo "preg_match('/(..).(..)/',s): {$a} {$b}, process time: {$duration}ms <br />";
?>

When I ran this on PhpFiddle Lite I got results like this:

substr(s,-2): 1 2, process time: 851ms
substr(s,3,2): 1 2, process time: 971ms
str_split(s,3): 1 2, process time: 1568ms
explode('-',s): 1 2, process time: 1670ms
preg_match('/(..).(..)/',s): 1 2, process time: 3328ms

The performance of substr with either (s, -2) or (s, 3, 2) as arguments perform almost equally well, provided you use only one call. Sometimes the second version came out as the winner. str_split and explode perform rather close, but not as well, and preg_match is the clear looser. The results depend on the server load, so you should try this on your own set-up. But it is certain that regular expressions have a heavy payload. Avoid them when you can do the job with the other string functions. I edited my answer when I realised that you can cast the original string immediately to int, which will ignore the part it cannot parse. This practically means you can get the first part as a number without calling any of the string functions. This was decisive to make substr the absolute winner!

Upvotes: 4

Payer Ahammed
Payer Ahammed

Reputation: 907

Try to convert the string to an array then use each array index to different variable your want

<?php
$str = '01-20'
$number = explode('-',$str);
$variable_1 = (int)$number[0];
$variable_2 = (int)$number[1];
?>

Upvotes: 0

Related Questions