Question

Like most of our code base, our mysql handling functions are custom built.

They work very well and include a number of logging forks.

A simplified version of our query execution function looks like this:

if(!$result=mysql_query($query)){
    file_put_contents(QUERYLOG,'Query '.$query.' failed execution');
}

This is overly simplified, but you get the basic idea: If queries fail, they will be logged to a separate query log.

This is a great way of keeping track of any queries that need to be looked at.

My question is as follows:

With the above, the tiny problem is that if a query fails both our query log, AND our php log will be stamped with the error as a mysql_query (… or mysql_connect, mysql_select_db, etc …) will produce a php error.

What we want to do is surpress the php error via:

.... $result=@mysql_query($query ....

So, as far as the question goes:

Does using the @ error suppression mechanism in php cause any performance impacts if no error is produced? Or does it only affect performance if an error is produced?

I know I know, micro optimization, but as you can guess, or query execution function is used millions of times a day, so even a small performance hit is worth examining.

Answer 1

made a little “research”

$s = microtime(true);
$a = array('1','2');
$b = $a[1];
echo microtime(true)-$s;

gives 1.1205673217773E-5
and if i use $b = @$a[1]; i get a bit more: 1.5974044799805E-5

so: yes, there is a difference, but no, you should not bother.