REGEXP not catching some names correctly if certain values are at certain positions in the string

I have the following regex meant to test against valid name formats:

^[a-zA-Z]+(([\'\,\.\- ][a-zA-Z ])?[a-zA-Z]*)*$

it seems to work fine with all the expected odd name possibilities, including the following:

o'Bannon
Smith, Jr.
Double-barreled
开发者_C百科

I'm having problem when I plug this into my PHP code. If the first character is a number it passes through as valid.

If the last character is a space, comma, full-stop or other special allowed character, it's failing as invalid.

My PHP code is :

$v = 'Tested Value';
$value = (filter_var($v, FILTER_VALIDATE_REGEXP,array("options"=>array("regexp"=>"^[a-zA-Z]+(([\'\,\.\-,\  ][a-zA-Z ])?[a-zA-Z]*)*$^"))));
if (strlen($value) <2 && strlen($v) !=0) {
    return "not valid";
}

What am I doing wrong here?

^[a-zA-Z]+(([\'\,\.\-,\  ][a-zA-Z ])?[a-zA-Z]*)*$^

The carets (^) at the beginning and end of the regex are being interpreted as regex deliminators, not as anchors. The regex isn't really matching the digits at the beginning of the string, it's skipping over them so it can start matching at the first letter it finds. You can use almost any ASCII punctuation character as the regex deliminator, but most people use # or ~, which are relatively uncommon and have no special meaning in regexes.

As for not allowing punctuation at the end, that's how the regex is written. Specifically, [\'\,\.\- ][a-zA-Z ] requires that each apostrophe, comma, period or hyphen be followed by a letter or a space. If you really want to allow any of those characters at the end, it's pretty simple:

~^(?:[a-z]+[',. -]*)+$~i

Of course, that's not a particularly good regex for validating names, but I have nothing better to offer; it's a job for which regexes are particularly ill-suited. And do you really want to be the one to tell your users their own names are invalid?

Your regex is way to complex

/^[a-z]+[',. a-z-]*$/i 

should do the same thing