Prevent non-latin characters

Murzim

Hello all,

I am validating a form input using preg_match:

if (preg_match("/[a-zA-Z0-9]/", $first_name)) {
echo 'accepted';
}
else{
echo 'not accepted';
}

Unfortunately this will also accept non-latin letters as well (such as Greek).

I want to prevent this but I am out of ideas.

Tried some mb functions with no luck.

Thank you before hand.

bpat1434

Sounds like you want to use the 0xFF type of notation for characters. I'm not sure what hex haracters you want are, but the English hex characters you want would be are 0x41-0x5A (A-Z), 0x61-0x7A (a-z), 0x30-0x39 (0-9).

I'm not saying it will eliminate greek characters; however, it's more accurate as to what characters you want. You can find all these hex, octal and other codes at http://www.asciitable.com or http://www.lookuptables.com

Murzim

Hello bpat1434,

Thank you for your answer. Unfortunately it doesn't seem to solve the issue with the non latin letters.

NogDog

You need to either anchor your pattern or invert it. Otherwise you are just checking that it contains at least one of those allowed characters.

// inverted:
if(preg_match('/[^a-zA-Z0-9]/', $first_name) == false)
{
   // accepted
}

// ...or... //

// anchored to start (^) and end ($) of string:
if(preg_match('/^[a-zA-Z0-9]$/', $first_name))
{
   // accepted
}

Murzim

Hello NogDog,

Thank you for your reply.

This:

// anchored to start (^) and end ($) of string:
if(preg_match('/^[a-zA-Z0-9]$/', $first_name))
{
   // accepted
}

has not worked.

This however:
// inverted:

if(preg_match('/[^a-zA-Z0-9]/', $first_name) == false)
{
   // accepted
}

worked smoothly!

I just added /t to accept white spaces: preg_match('/[^a-zA-Z0-9 /t]/', $first_name)

Thank you very much!

Weedpacket

// anchored to start (^) and end ($) of string:
if(preg_match('/^[a-zA-Z0-9]$/', $first_name))

Typo. Correction:

// anchored to start (^) and end ($) of string:
if(preg_match('/^[a-zA-Z0-9]+$/', $first_name))

Murzim

whats the difference if you add the + part ?

bpat1434

+, and {X,Y} are quantifiers. meaning 0 or more, + meaning 1 or more and {X,Y} meaning at least X and at most Y occurances.

For example, given this string:

Abbott and Costello Meet Dr. Jekyll and Mr. Hyde

Looking for say "Abb" you could either look for abb or a(b)+ or a(b){2}. Each is perfectly acceptable; however, each has it's advantages and disadvantages.

The first would only match those that contain Abb exactly (no more or less). The third would do the same; however, there is some lee-way because you can define how many "b"s you want. By not passing a second parameter, you're saying I only want an "a" followed by 2 "b"s. The second one would match Abb as well as Abbbbbbbbbbbbb.

You can take these rules and then apply them to some more intricate things like IP address:

^[\d]{1,3}\.[\d]{1,3}\.[\d]{1,3}\.[\d]{1,3}$

-- OR --

^([\d]{1,3}){4}$

Or for email addresses:

^[a-z0-9]+[a-z0-9_.-]*@[a-z0-9]+\.[a-z0-9]{2,4}(\.[a-z0-9]{2,4})?

So with the quantifiers you can really leverage the regex to match (or not match) exactly what you're looking for.

For more information, you can look at www.regular-expressions.info for some good information. You can really do some intersting stuff with positive & negative look ahead/behinds and conditional expressions.