PHP strlen(): String Length in Bytes and Characters

strlen($string) returns the length of a string in bytes. For characters, which is what you want for Japanese text, accented letters and emoji, use mb_strlen($string). For plain English text the two give the same number.

<?php
echo strlen('Hello'), "\n";         // 5 bytes
echo mb_strlen('Hello'), "\n";      // 5 characters

echo strlen('こんにちは'), "\n";     // 15 bytes
echo mb_strlen('こんにちは'), "\n";  // 5 characters

Spaces, punctuation and newlines all count. strlen("a b\n") is 4.

Bytes vs characters in UTF-8

PHP strings are sequences of bytes, and UTF-8 uses a different number of bytes per character: 1 for ASCII letters and digits, 2 for letters such as é or ж, 3 for kana, common kanji, Chinese characters and symbols such as €, and 4 for most emoji and rare kanji such as 𠮷. strlen() adds up the bytes; mb_strlen() counts the characters.

<?php
$samples = ['A', 'é', 'ж', 'あ', '漢', '€', '😀'];

foreach ($samples as $char) {
    printf("%s  strlen=%d  mb_strlen=%d\n", $char, strlen($char), mb_strlen($char));
}

$mixed = 'PHPで文字数';
printf("%s  strlen=%d  mb_strlen=%d\n", $mixed, strlen($mixed), mb_strlen($mixed));

Use strlen() when bytes matter: a database column limited in bytes, a file size, a binary protocol. Use mb_strlen() for anything a person reads: a character counter, a "max 20 characters" rule, cutting text with mb_substr().

Emoji and characters made of several code points

mb_strlen() counts Unicode code points. Some things a reader sees as one character are several code points joined together: a family emoji, a flag, a skin tone, or a letter plus a combining accent.

<?php
$items = [
    'thumbs up'      => '👍',
    'with skin tone' => '👍🏽',
    'flag of Japan'  => '🇯🇵',
    'family'         => '👨‍👩‍👧',
    'e + accent'     => "e\u{0301}",
];

foreach ($items as $name => $text) {
    printf("%-15s %s  strlen=%2d  mb_strlen=%d\n", $name, $text, strlen($text), mb_strlen($text));
}

Counting what a reader sees as single characters (grapheme clusters) needs grapheme_strlen() from the intl extension:

echo grapheme_strlen('👨‍👩‍👧');   // 1
echo grapheme_strlen('🇯🇵');     // 1

For a character counter on a Japanese form, mb_strlen() is usually enough: every kana and kanji is a single code point.

Display width with mb_strwidth

On a terminal or in a fixed-width layout, a full-width Japanese character takes two columns while a half-width katakana or Latin letter takes one. mb_strwidth() returns that width, which is what Japanese "全角は2文字" rules usually mean.

<?php
foreach (['abc', 'ｱｲｳ', 'アイウ', '日本語', 'PHP入門'] as $s) {
    printf("%s  mb_strlen=%d  mb_strwidth=%d\n", $s, mb_strlen($s), mb_strwidth($s));
}

Check if a string is empty

The clearest test is a strict comparison with ''. strlen($s) === 0 gives the same answer. empty($s) is not the same: it also returns true for the string "0", which is a valid answer in many forms.

<?php
foreach (['', '0', ' ', 'a'] as $s) {
    printf(
        "%-5s  === '': %-5s  strlen: %d  empty(): %s\n",
        var_export($s, true),
        var_export($s === '', true),
        strlen($s),
        var_export(empty($s), true)
    );
}

A string of spaces is not empty. Run it through trim() first if spaces alone should count as nothing.

strlen(null) and non-strings

strlen() expects a string. Passing null still returns 0, but since PHP 8.1 it prints a deprecation notice:

Deprecated: strlen(): Passing null to parameter #1 ($string) of type string is deprecated in /home/index.php on line 2

An array throws TypeError: strlen(): Argument #1 ($string) must be of type string, array given.

Cast or default the value first, for example strlen($value ?? ''). Integers are converted to strings, so strlen(12345) is 5. To count array elements, use count().

Validate the length of user input

A typical rule: a username of 3 to 16 characters, after trimming. Use mb_strlen() so a Japanese name is measured in characters, not bytes.

<?php
function checkUsername(string $name): string
{
    $name = trim($name);
    $length = mb_strlen($name);

    if ($length < 3) {
        return "\"$name\" is too short ($length characters, min 3)";
    }
    if ($length > 16) {
        return "\"$name\" is too long ($length characters, max 16)";
    }
    return "\"$name\" is OK ($length characters, " . strlen($name) . ' bytes)';
}

echo checkUsername('ad'), "\n";
echo checkUsername('ada_lovelace'), "\n";
echo checkUsername('やまだたろう'), "\n";
echo checkUsername('a_very_long_username_here'), "\n";

With strlen() in place of mb_strlen(), やまだたろう would count as 18 and be rejected as too long, even though it is six characters.