Validating a prename and surname

Question

I want to validate a prename and surname for users to register at a worldwide social community.

I want to allow all these language-related special characters like ÄÖÜÀýéè and a maximum of 3 prenames, all seperated with a space (if there are more than 1 prename). Also no underscores should be allowed or other special chars, but minus (-) should be allowed.

That's all I can think of what should be allowed in a name. I don't know how it is in other countries than Germany, but here you can have up to 3 prenames and a surname can also have 2 words like "von Seidenfeld". Maybe you can also give me some more suggestions here, because I want to cover this name-validation-system for every full name around the world.

I also know I've to use RegEx, but I don't know how to allow only these conditions that I just wrote.

What I have so far:

if (!preg_match('/^(\pL+\s+)$/u', $value)) echo 'error';

or view here: http://regex101.com/r/tE0uQ5 (sadly doesn't work, no matches)

Answer 1

That's all I can think of what should be allowed in a name.

Not even close.

I don't know how it is in other countries than Germany, but here you can have up to 3 prenames and a surname can also have 2 words like "von Seidenfeld".

It's different in other countries. It's probably also different in Germany, just not for your name or the people's names that you're immediately thinking of right this moment.

I want to cover this name-validation-system for every full name around the world.

This should be required reading on the subject. But to summarize... There is no way to do this. None. If you absolutely need a regular expression that will match a name, this might help.

By definition, anything that somebody enters into your system as their name is their name, at least as far as your system is concerned. It's how they've decided you should identify them. Is it their legal name in their jurisdiction? In your jurisdiction? Is it the name their parents call them? The name their friends call them? None of these things can be validated by a regular expression.

In short... Just accept any non-empty input as a name.

Answer 2

Forget name validation if it's international. Have an example, this is a totally valid Hungarian name:

dr. Lakatosné Dr. Bíró-Kis Imola

Imola is the given name, the rest is family name just to mix up things a bit more. Good luck validating that against anything. If you want more headache-inducing examples, take the artists, who are for example legally allowed to register an artist name, which becomes part of their official name.

Generally speaking, names have different origins, some cultures for example had the habit of always adding the father's name as well as appending an own name. The longest name I have heard of in my family is 7 (!) given names, not to mention various prefixes and honorary titles, etc. that some people insist on using everywhere they can and which is in fact a part of their official name (appears on their ID card). To name a famous example for a really long name, check out Rudolph Valentino, who had a total of 9 names.

If you want to build a truly international system, you'll create two fields, family and given name, and let the user enter whatever they want. When displaying, be sure to take the name order of the person viewing the site into account. (Hungarian is switched for example.) Also, take great care to require input in only one of these fields because some people may not have a family name.

Answer 3

I'd suggest looking for invalid characters rather than valid ones, to simplify things:

^((?:(?:(?<!^)[ ])?[^*&@_#$\]\\\^\d ]+){1,3})$

I had to escape the 4 special characters at the end (], \, ^, -). Also added \d to filter out numbers.

See it in action here! http://regex101.com/r/vY9wW5

I'm betting no one cares anymore, but I got this puppy working. The problem was me realizing how PHP strings should be escaped (with single quotes, just the escaped backslashes needed to be doubled up) and, the regex testers and maybe online ides were messing with my string in subtle ways.

<?php
$sourcestring="Uwe\nUwe-Jens\nGünther Kalle\nKalle Jens Chantal";
$regex = '((?:(?:(?<!^)[ ]|(?<=(?<=^|\n)))[^*&@_#$\]\\\\\^\d \n]+){1,3}(?=\n|$))';
preg_match_all($regex,$sourcestring,$matches);
echo "<pre>".print_r($matches,true);
?>

That said, I think there are cases for names with weird symbols, such as numbers and symbols that are rare enough where you can make them work around a reality-check filter, in exchange for being able to filter out the button smashers from your login process.