&quot;a $b&quot;; vs. 'a'.$b;

"a $b"; vs. 'a'.$b;

Anon

I find that many people don't concatenate enough. Personally I've recently become a concatenating fanatic but am curious how others feel on this topic. Here is an example of what I mean :

echo "<a href=\"$link\">$link</a>\n";

vs.

echo '<a href="'.$link.'">'.$link.'</a>'."\n";

Sure the second is longer but imho much more fun 😉 It's also faster and less obtrusive on the server (PHP doesn't parse anything within ''). So actually, this isn't a question but more of an observation. Please people, concatenate your strings! Once you get the hang of it it's actually easier, doing the old \" is annoying, right?

Also, the most useful string tutorial I've found is this :

http://www.zend.com/zend/tut/using-strings.php

Working with strings is a very underrated art yet it's what we do most. /rant.

Anon

Well, Philip, I'm going to be contrary and argue for the other side.

It remains to be proved that the overhead of parsing a double-quoted string is significant. I haven't read that part of the Zend PHP engine, but I do know that running through a C string with a pointer can compile into some honkin' fast machine code. (I used to play around with that sort of thing in the 1980s when I wrote a Usenet news handler for my Atari ST.)

My time, my sanity, and the readability of my code are important to me. I don't want to be dinking around with the concatenation operator, jumping in and out of quoted strings, counting the quotes to make sure I'm coming out even, and debugging the results unless I'm pretty darned sure that this is a legitimate performance issue.

And I've seen enough cases where theoretical advantages turned out to be either nonexistent or false to say("Show me the $benchmarks!")

Anon

Very interesting. Originally this post was a question of 'is it that much faster?' but ended up changing the focus midstream. Will have to create a file or two to test this out once and for all! Regarding counting quotes, haha yes that's true! Same though for counting backslashes. Point well taken though, \" is easier on the eyes then let's say '" .

Perhaps the latest phpbuilder article will be helpful in this quest.

Anon

The speed difference is likely to exist, but be so darned negligible (?) that it would be hard to spot. I feel there's much to be said for concatenation though. Not least ambiguity. eg.

$a = "$ba";

When what you actually wanted:

$a = $b.'a';

Personally I feel the former is less "clean". Plus the colors on editor show the latter more clearly. 🙂

Anon

If ease of coding is what you are after I prefer sprintf and printf I really don't know what issue lies with performance and using these but I find them a lot cleaner and easier to manage:

$string = sprintf('<a href="%s">click here</a>', $PHP_SELF);

And plus you don't have to worry about concat or escaping quotes, just my two cents, and if anyone know s anything about this being terribly slower please let me know, for I tend to use it alot

Anon

I've not a clue how much slower it is but as you know, they do formatting on the string so this takes extra time. There is an article regarding 'top 21 programming mistakes' that refers to overusing printf() here :

www.zend.com/zend/art/mistake.php

Don't worry though, it is the #21 mistake and not #1 🙂

Anon

Thanks Philip I feel a whole lot better now... ha ha. I use that damn printf() everywhere because it helps me keep my code organized, I am just wondering if it is goign to make that much of a difference, I am not talking about millions of hits a day so, I think I am safe

Anon

OK, guys. I dragged out ab, the Apache Benchmark tool, and ran it against three sets of code.

inline substitution,

<?
$str1 = "The quick brown fox";
$str2 = "jumped over";
$str3 = "the lazy dog's back";

for ($i=0; $i<1000; $i++)
print("$str1 $str2 $str3");

printf

<?
$str1 = "The quick brown fox";
$str2 = "jumped over";
$str3 = "the lazy dog's back";

for ($i=0; $i<1000; $i++)
printf("%s %s %s", $str1, $str2, $str3);

concatenation

<?
$str1 = "The quick brown fox";
$str2 = "jumped over";
$str3 = "the lazy dog's back";

for ($i=0; $i<1000; $i++)
print($str1 . " " . $str2 . " " . $str3);

For each of these files, I ran 1,000 retrievals using ab's default settings, localhost (no Ethernet traffic contention). That's 10,000 invocations of each technique.

printf 47.830 seconds

inline 49.019 seconds
concat 42.126 seconds

I got similar results with multiple runs.

So, it's up to you whether you regard these differences as significant. I don't think it's worth worrying about, and I'll use the style that I find the to be the most legible and maintainable.

Anon

Oops.

Of course, 1000 * 1000 is not 10,000. Don't let me near a checkbook. :-)

Anon

very cool but i have a question, regarding this :

print($str1 . " " . $str2 . " " . $str3);

shouldn't it be :

print($str1 . ' ' . $str2 . ' ' . $str3);

as it seems the " " would still be parsed. try it with ' ' , i'm curious to see if there is a difference.

philip

Anon

You're right. The difference across a million lines is pretty small.

41.495 with " " changed to ' '.

Anon

I'll use the style that I find the to be the most legible and maintainable.

Hear, hear! It's about time to trot out Knuth's dictum:

Premature optimization is the root of all evil.

Kernighan and Plauger (in Software Tools, I think) likewise
inveigh against blindly assuming that one needs to worry about
efficiency before readability, and the need for genuine metrics.
So thanks for taking the time to perform this little benchmark!

Anon

Premature optimization is the root of all evil.

All things in moderation though of course. Foregoing thinking about how you write your scripts and what functions to use (eg. ereg_replace instead of str_replace) means you will quite likely end up with much slower scripts. And should these scripts be put on a site that gets a significant amount of traffic, then the difference will sure mount up.