Variable Interpolation

douceur

Since echo 'var is $var' and echo "var is $var" produce different outputs, I have a question. In situations where no variables need to be interpolated, is it actually faster to use single quotes? Or does it not really matter?

Basically, is it faster to use echo 'hello world' or echo "hello world"? Or are both equally fast?

Sxooter

Benchmark it. My guess is that while one is probably a bit faster, it's not imporrtant compared to say, one regex operation.

Sxooter

I ran this code:

#!/usr/local/bin/php -q
<?php
$max = 1000000;
for ($i=0;$i<$max;$i++){
        print "Hello";
}
?>

with this command:
time ./k >/dev/null
then changed the ""s to ''s and ran it again. The difference was too small to measure.

douceur

Strange... I ran a similar test, but came up with very different results. I used the code below, except that in the real code, the string contained Hello World written 350+ times, not just once. I ran it three times with "s, and three times with 's.

<?php
$timeparts = explode(" ",microtime());
$starttime = $timeparts[1].substr($timeparts[0],1);
$max = 1000000;
for ($i=0; $i < $max; $i++){
	$string = "Hello World";
}
$timeparts = explode(" ",microtime());
$endtime = $timeparts[1].substr($timeparts[0],1);
echo bcsub($endtime,$starttime,6)
?>

Results were:

" = 8.483677, 7.578489, 8.085688
' = 12.464859, 12.479967, 12.478757

I suppose that, seeing as it looped 1000000 times, the difference is actually pretty miniscule. 4.5 seconds divided by 1000000 is pretty darn small! But either way, I definitely would have thought that the results would have been the other way around.

laserlight

From what I've heard, they should not make much of a difference.
The difference only comes when variables are parsed.

I tested for output with:

<?php
$start_time = microtime();
$max = 50000;
for ($i = 0; $i < $max; $i++){
    echo 'Hello World<br>';
}
$end_time = microtime();
$start_array = explode(" ", $start_time);
$end_array = explode(" ", $end_time);
$start_time = array_sum($start_array);
$end_time = array_sum($end_array);
echo '<p>Start time: ' . $start_time . '<br>';
echo 'End time: ' . $end_time . '<br>';
echo 'Time diff: ' . ($end_time - $start_time) . '</p>';
?>

My results were:
$max = 10000
" 4.5339829921722
' 4.018935084343

$max = 100000
" 27.598752975464
' 30.377369999886

$max = 50000
" 17.05751490593
' 14.133512020111

$max = 150000
" 28.014575004578
' 29.235834002495

From that I can conclude:
a) my benchmarking script is quite poor
b) server resources vary
c) it really doesnt matter if no variables are parsed within the double quotes.

I would be tempted to conclude that as $max tends to infinity double quotes are preferable to single quotes, but with only 4 tests on a single remote comp that is an absurd claim.

douceur

Well it seemed to me that the main difference occurred when the string was rather large.

Weedpacket

Originally posted by laserlight
From what I've heard, they should not make much of a difference.
The difference only comes when variables are parsed.

That's the experience I had; as soon as variables appeared things slowed right down.

Originally posted by doceur
Well it seemed to me that the main difference occurred when the string was rather large.

If you're echoing, then of course the thing to do then is break out of PHP completely for the duration 🙂. If you're shoving stuff into a variable... dunno, never tried: would output buffering speed things up (turn OB on, break out of PHP, do the text, reenter PHP, grab and clear the bufer's contents)? Heredoc syntax does variable interpolation, so I don't see it being recognisable faster than double-quotes as a first guess.