No need to use isset and empty in the condition

[roman]

PHP's empty() returns true if its argument isn't initialized. So, for example, doing the following is redundant and wastes time:

isset($data['comments']) && !empty($data['comments'])

[leng]

roman, this will throw a notice in PHP if $data['comments'] is never defined before the use of empty($data['comments']

This will create a Notice

Code: Select all

//it's never defined anywhere before
if(!empty($data['comments'])){

}

This is perfect E_STRICT compliant

Code: Select all

if( isset($data['comments']) && !empty($data['comments']) ){

}

Having the conditional order does affect PHP too, NOTICE!

Code: Select all

if( !empty($data['comments']) && isset($data['comments']) ){

}

[roman]

Leng, please help me reproduce what you think should happen. Here's the code I am using to test this:

Code: Select all

<?php
error_reporting(E_ALL | E_STRICT);
echo empty($data['message']);
?>

All I get is 1.

I use PHP 5.2.9 on Windows Vista with XAMPP.

[leng]

hey yea right. I was wrong.
What I said should be for other functions than empty()

[leng]

btw the following is not completely redundant and wastes time

Code: Select all

isset($data['comments']) && !empty($data['comments'])

isset is faster than empty (well isset only checks if the var is set or not)
and if the first condition is false (isset), the second (empty) won't be checked, because it's an AND &&
So you have a faster conditional logic here (if most of the time you know that $data['comments'] will not be set at all)

[roman]

I know about the order of evaluation of conditions in PHP. It's the same in C.

I agree that if most of the time you know that $data['comments'] is not be set, then doing both may be OK. However, is the difference in speed of those functions great enough to justify increasing amount of code and probability of an error? Maybe empty() does the same thing that isset() does at first? Have you tested the speed of both of them when the argument is not set?

[leng]

No I didn't test personally but I think I have seen some benchmark some where...