Re: 1 last error to fix before the application is done!

Colin Guthrie <gmane@xxxxxxxxxxxxxx> · Tue, 14 Oct 2008 14:51:00 +0100

Yeti wrote:
Personally, I very rarely see the point in using array_key_exists... It's a function call and has overhead where as isset() and empty() are language constructs and (I would hope) are much more efficient (although I've not done any benchmarks)

# i don't know what's wrong with this ..

$foo = array();
$foo['bar'] = null;
if (array_key_exists('bar', $foo)) {
 switch ($foo['bar']) {
  case true:
   //do something
   break 1;
  default:
   //do something else
}
}

# if that's causing more overhead then let me see your benchmark, then
i believe it

I don't really get what you're trying to say here.

All I'm saying is that empty(), as a language construct, should be more 
efficient than the resulting function call overhead when using 
array_key_exists().

As I highlighted in my previous example, the only way to test of an 
array has an index which is set to null is to use the array_key_exists() 
function as both isset() (also a language construct) and empty() will 
not be able to tell you this.

So your example above (which sets an array index to null) is the only 
way to do what you want to do (the switch() is overkill for an if/else 
scenario but that's beside the point).

All I was saying was that for an array index value that is set to a 
known true/false value, or not set at all, using !empty() to test for 
when the value is true is the most efficient way.

Consider the following two attached test cases to check for a true 
value, two indexes deep in the array. One uses two calls to 
array_key_exits and then a comparison, the other uses !empty(). They are 
functionally equivalent but the results speak for themselves:

[colin@jimmy 0.11]$ time php test1.php
Done: 10000000
13.76user 0.01system 0:14.08elapsed 97%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2805minor)pagefaults 0swaps
[colin@jimmy 0.11]$ time php test2.php
Done: 10000000
3.06user 0.03system 0:03.13elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2797minor)pagefaults 0swaps

And a repeat for random variations:

[colin@jimmy 0.11]$ time php test1.php
Done: 10000000
14.58user 0.05system 0:14.77elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2805minor)pagefaults 0swaps
[colin@jimmy 0.11]$ time php test2.php
Done: 10000000
3.01user 0.01system 0:03.06elapsed 98%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2797minor)pagefaults 0swaps

There's a benchmark for you (not sure if it's the one you wanted) which 
proves that the !empty() is ~4.5x faster.

 Use !empty() :p

Col

--

Colin Guthrie
gmane(at)colin.guthr.ie
http://colin.guthr.ie/

Day Job:
  Tribalogic Limited [http://www.tribalogic.net/]
Open Source:
  Mandriva Linux Contributor [http://www.mandriva.com/]
  PulseAudio Hacker [http://www.pulseaudio.org/]
  Trac Hacker [http://trac.edgewall.org/]
Attachment:
test1.php

Description: application/php
Attachment:
test2.php

Description: application/php
-- 
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php