PHP - optimizing multiple array_merge ()

A small hint that I am going to talk about now, I learned for quite some time. But, since I just discovered that it was in this place that optimization helped to save a fair amount of time already in the current project, I decided to share it with the public (and the men don’t know, yeah).

To be brief, here is a test script and what it outputs:

$ start = microtime (true);
$ res1 = array ();
for ($ i = 0; $ i <1000; $ i ++) {
        $ res1 = array_merge ($ res1, array (1, 2, 3));
}
echo "1000 merges:". ceil ((microtime (true) - $ start) * 1000). "ms \ n";
 
$ start = microtime (true);
$ toMerge = array ();
for ($ i = 0; $ i <1000; $ i ++) {
        $ toMerge [] = array (1, 2, 3); 
}
$ res2 = call_user_func_array ('array_merge', $ toMerge);
echo "call_user_func_array ('array_merge', ..):". ceil ((microtime (true) - $ start) * 1000). "ms \ n";
 
echo "Is it true that the two arrays are equal? ​​It is";
var_export ($ res2 === $ res1);
echo ". \ n";

~ % php ~/tmp/array_merge.php
1000 merges: 980ms
call_user_func_array('array_merge',..): 11ms
Is it true that the two arrays are equal? It is true.

The savings are substantial. Maybe someone will come in handy.