Using preg_replace to make ordinals appear in all lower case

Question

I have a column of address data that has ordinals displayed like this:

3Rd Floor, Cumbrian House

Room 223, 2Nd Floor

when I would like them to be displayed like this instead:

3rd Floor, Cumbrian House

Room 223, 2nd Floor

I'm trying to make use of the preg_replace function to swap out a capitalised letter, that directly follows a number, with a lower case letter instead. (I admit this is my first time using the preg_replace function. I am however fine with preg_match and regular expressions).

So far I have:

$string = '3Rd Floor, Cumbrian House';

$ordinalregex = '/(^.*\d+)([A-Z])/';
$correctordinal = '$1'.strtolower('$2');

echo preg_replace($ordinalregex,$correctordinal,$string)."<br>";

But its not having the desired effect and outputs the line exactly as it first appeared.

Thanks

Answer 1

I think you need to use preg_replace_callback as I'm not sure you can replace based on the part you matches in normal preg_replace. What you need to do is use a callback function that takes the matches array and processes it (you could also just use preg_match and then work with the matches seperetaly).

This example works (although it uses an anonymous function for the callback, which requires PHP 5.3.0+):

function fixOrdinals($string) { 
    return preg_replace_callback("/[0-9][A-Z]/",
        function ($matches) { 
            return strtolower($matches[0]);
        },
        $string
    );
}

echo fixOrdinals("3Rd Floor, Cumbrian House")."\n";
echo fixOrdinals("Room 223, 2Nd Floor")."\n";
echo fixOrdinals("4Th Room, 2Nd Floor")."\n";

Output:

3rd Floor, Cumbrian House
Room 223, 2nd Floor
4th Room, 2nd Floor

Answer 2

<?php

    $ordinalregex = '/(\d+)(st|nd|rd|th)/ei';
    $input = "Room 223, 2Nd Floor";

    $input = preg_replace($ordinalregex, "strtolower('$1$2')", $input);

    print($input);

The /e modifier tells the interpreter to use eval. Now this works, however it's NOT SECURE. If your input has a \0 (null char) due to a C bug the interpreter will execute the part of the sting after the \0.