Since someone mentioned efficiency, I've benchmarked the functions given so far out of curiosity:
function startsWith1($str, $char) {
return strpos($str, $char) === 0;
}
function startsWith2($str, $char) {
return stripos($str, $char) === 0;
}
function startsWith3($str, $char) {
return substr($str, 0, 1) === $char;
}
function startsWith4($str, $char){
return $str[0] === $char;
}
function startsWith5($str, $char){
return (bool) preg_match('/^' . $char . '/', $str);
}
function startsWith6($str, $char) {
if (is_null($encoding)) $encoding = mb_internal_encoding();
return mb_substr($str, 0, mb_strlen($char, $encoding), $encoding) === $char;
}
Here are the results on my average DualCore machine with 100.000 runs each
// Testing '_string'
startsWith1 took 0.385906934738
startsWith2 took 0.457293987274
startsWith3 took 0.412894964218
startsWith4 took 0.366240024567 <-- fastest
startsWith5 took 0.642996072769
startsWith6 took 1.39859509468
// Tested "string"
startsWith1 took 0.384965896606
startsWith2 took 0.445554971695
startsWith3 took 0.42377281189
startsWith4 took 0.373164176941 <-- fastest
startsWith5 took 0.630424022675
startsWith6 took 1.40699005127
// Tested 1000 char random string [a-z0-9]
startsWith1 took 0.430691003799
startsWith2 took 4.447286129
startsWith3 took 0.413349866867
startsWith4 took 0.368592977524 <-- fastest
startsWith5 took 0.627470016479
startsWith6 took 1.40957403183
// Tested 1000 char random string [a-z0-9] with '_' prefix
startsWith1 took 0.384054899216
startsWith2 took 4.41522812843
startsWith3 took 0.408898115158
startsWith4 took 0.363884925842 <-- fastest
startsWith5 took 0.638479948044
startsWith6 took 1.41304707527
As you can see, treating the haystack as array to find out the char at the first position is always the fastest solution. It is also always performing at equal speed, regardless of string length. Using strpos is faster than substr for short strings but slower for long strings, when the string does not start with the prefix. The difference is irrelevant though. stripos is incredibly slow with long strings. preg_match performs mostly the same regardless of string length, but is only mediocre in speed. The mb_substr solution performs worst, while probably being more reliable though.
Given that these numbers are for 100.000 runs, it should be obvious that we are talking about 0.0000x seconds per call. Picking one over the other for efficiency is a worthless micro-optimization, unless your app is doing startsWith checking for a living.
