Check if a String is mostly ALL CAPS in PHP

user_24224 0
Viewed 69

What is the best way to see if a string contains mostly capital letters?

The string may also contain symbols, spaces, numbers, so would still want it to return true in those cases.

For example: I can check if a strings is ALL-CAPS by something similar to this:

if (strtoupper($str) == $str) { /* its true */ }

But what if we need to determine if a string is 80% or more ALL-CAPs.

THE 15 SMALL BROWN FOXES JUMP INTO THE BURNING barn! -> true

The 15 Small Brown Foxes JUMP Into the Burning Barn! -> false

I can loop though all the characters, checking them individually, but thats seems a bit wasteful imho.

Is there a better way?

php string
4 Answers
$countUppercase = strlen(preg_replace('/[^A-Z]+/', '', $str));
// or: mb_strlen(...)

... and then divide by strlen($str)

user_172184290

A simple for loop should give the best performance

$numUpper = 0;

for ($i = 0; $i < strlen($str); $i++){
    if (ctype_upper($str[$i])) {
        $numUpper++;
    }
}

return $numUpper;
user_118087880

Another option could be using preg_match_all which returns the number of full pattern matches and mb_strlen.

The pattern \p{Lu} matches an uppercase letter that has a lowercase variant.

For example:

function mostlyUpperInString($s, $threshold) {
    return preg_match_all("/\p{Lu}/u", $s) / mb_strlen($s) > $threshold;
}

function moreUpperThanLower($s, $threshold) {
    return preg_match_all("/\p{Lu}/u", $s) / preg_match_all("/\P{Lu}/u", $s) > $threshold;
}

$strings = [
    "THE 15 SMALL BROWN FOXES JUMP INTO THE BURNING barn!",
    "The 15 Small Brown Foxes JUMP Into the Burning Barn!"
];

foreach ($strings as $str) {
    echo $str . " -> 80% mostlyUpperInString: ". (mostlyUpperInString($str, 0.8) ? "true" : "false") . PHP_EOL;
    echo $str . " -> 80% moreUpperThanLower: ". (moreUpperThanLower($str, 0.8) ? "true" : "false") . PHP_EOL;
    echo PHP_EOL;
}

Output

THE 15 SMALL BROWN FOXES JUMP INTO THE BURNING barn! -> 80% mostlyUpperInString: false
THE 15 SMALL BROWN FOXES JUMP INTO THE BURNING barn! -> 80% moreUpperThanLower: true

The 15 Small Brown Foxes JUMP Into the Burning Barn! -> 80% mostlyUpperInString: false
The 15 Small Brown Foxes JUMP Into the Burning Barn! -> 80% moreUpperThanLower: false

See a PHP demo.

user_54249880

I wanted to play around with alternative approaches and came to something like this. Honestly I'd probably just write a loop. Just more clear than something like this and you can omit numbers and other things counted towards "lower case".

Though, with a little massaging of your input you might be able to do something along these lines, simply comparing differences between two strings.


<?php
   
isMostlyUpperCase('THE 15 SMALL BROWN FOXES JUMP INTO THE BURNING barn!');
isMostlyUpperCase('The 15 Small Brown Foxes JUMP Into the Burning Barn!');
   
function isMostlyUpperCase($strIn) {
  $strAsUpper = str_split(strtolower($strIn));
  $str = str_split($strIn);
  $diff = array_diff_assoc($str, $strAsUpper);
  return (count($diff) > strlen($strIn) - count($diff)); 
}

I've added functional examples based on the other two answers:

<?php
    
$str1 = 'THE 15 SMALL BROWN FOXES JUMP INTO THE BURNING barn!';
$str2 = 'The 15 Small Brown Foxes JUMP Into the Burning Barn!';

function pregIsMostlyUpper($str)
{
    $countUppercase = strlen(preg_replace('/[^A-Z]+/', '', $str));
    $countLowercase = strlen(preg_replace('/[^a-z]+/', '', $str));
    return (bool)($countUppercase > $countLowercase);
}

function loopIsMostlyUpper($str)
{
    $numUpper = 0;
    $numLower = 0;
    for ($i = 0;$i < strlen($str);$i++)
    {
        if (ctype_alpha($str[$i]))
        {
            if (ctype_upper($str[$i]))
            {
                $numUpper++;
            } else {
                $numLower++;
            }
        }
    }
    return (bool)($numUpper > $numLower);
}

var_dump(loopIsMostlyUpper($str1));  //true
var_dump(loopIsMostlyUpper($str2));  //false
var_dump(pregIsMostlyUpper($str1));  //true
var_dump(pregIsMostlyUpper($str2));  //false

I'll let you time them if you want! Try and beat a sed / awk solution.

user_13038680
Related