PHP substr(): Get Part of a String

substr($string, $offset, $length) returns the part of $string that starts at position $offset (counted from 0) and is $length characters long. substr('Hello world', 0, 5) returns "Hello". Leave out $length to get everything to the end.

<?php
$s = 'Hello world';

echo substr($s, 0, 5), "\n";   // from the start, 5 characters
echo substr($s, 6), "\n";      // from position 6 to the end
echo substr($s, 4, 3), "\n";   // 3 characters from position 4

substr() counts bytes. That is the same as characters for English text, but not for Japanese, Chinese, emoji or accented letters. For those use mb_substr(), shown below.

Syntax

substr(string $string, int $offset, ?int $length = null): string

• $offset is where to start. 0 is the first character. A negative value counts from the end.
• $length is how many bytes to take (one byte is one character in plain ASCII text). A negative value leaves that many off the end. null (or leaving it out) means "to the end".

Negative offset and length: work from the end

A negative $offset starts counting from the end of the string, so -1 is the last character. A negative $length stops that many characters before the end.

<?php
$file = 'report-2026.pdf';

echo substr($file, -3), "\n";       // last 3 characters
echo substr($file, -8, 4), "\n";    // 4 characters, starting 8 from the end
echo substr($file, 0, -4), "\n";    // everything except the last 4
echo substr($file, 7, -4), "\n";    // from position 7, minus the last 4

substr($s, 0, -1) is the usual way to remove the last character, for example a trailing comma left by a loop.

Get the first or last character

For a single byte you can also index the string directly. $s[0] is the first character and, since PHP 7.1, $s[-1] is the last.

<?php
$word = 'banana';

echo substr($word, 0, 1), ' ', $word[0], "\n";    // first character
echo substr($word, -1), ' ', $word[-1], "\n";     // last character
echo substr($word, 1), "\n";                     // drop the first character

Indexing a position that does not exist prints a warning, while substr() just returns an empty string. If the string might be empty, prefer substr().

Out of range: an empty string, not false

If $offset is past the end of the string, substr() returns "" in PHP 8. In PHP 7 it returned false, so old code that checks === false no longer works.

<?php
var_dump(substr('abc', 5));       // start past the end
var_dump(substr('abc', 1, 100));  // length longer than the rest is fine
var_dump(substr('abc', -10));     // negative offset before the start: from 0

Japanese text: use mb_substr

UTF-8 stores kana and common kanji in 3 bytes each (rare kanji such as 𠮷 and emoji take 4). substr() counts bytes, so substr('こんにちは', 0, 3) returns only こ, and a length that does not land on a character boundary cuts a character in half and produces broken text. mb_substr() takes the same arguments and counts characters.

<?php
$greeting = 'こんにちは世界';

echo substr($greeting, 0, 3), "\n";       // 3 bytes = 1 character
echo mb_substr($greeting, 0, 3), "\n";    // 3 characters
echo mb_substr($greeting, -2), "\n";      // last 2 characters

// Cutting at 4 bytes splits a character: the result is not valid UTF-8
var_dump(mb_check_encoding(substr($greeting, 0, 4), 'UTF-8'));

The same applies to emoji and accented letters such as é. Character counts use mb_strlen(); see strlen() for bytes vs characters.

Shorten text with an ellipsis

Cutting a long title for a list or card is the most common real use of substr(). Use mb_substr() so a Japanese title is never cut mid-character, and only add ... when you actually shortened something.

<?php
function excerpt(string $text, int $max): string
{
    if (mb_strlen($text) <= $max) {
        return $text;
    }
    return rtrim(mb_substr($text, 0, $max)) . '...';
}

echo excerpt('Short title', 20), "\n";
echo excerpt('A much longer title that will not fit in the card', 20), "\n";
echo excerpt('東京から大阪までの新幹線の旅', 8), "\n";

mb_strimwidth($text, 0, $width, '...') does the same by display width, where a full-width Japanese character counts as 2 columns and a Latin letter as 1. It suits fixed-width layouts.

Count a substring with substr_count

substr_count($haystack, $needle) returns how many times $needle appears. It is case-sensitive and does not count overlapping matches.

<?php
echo substr_count('banana', 'an'), "\n";
echo substr_count('aaaa', 'aa'), "\n";      // 2, not 3: no overlaps
echo substr_count('Hello hello', 'hello'), "\n";
echo substr_count(strtolower('Hello hello'), 'hello'), "\n";

substr with strpos: text before or after a character

Combine substr() with strpos() to cut at a character whose position you do not know in advance, such as the @ in an email address:

<?php
$email = 'ada.lovelace@example.com';
$at = strpos($email, '@');

echo substr($email, 0, $at), "\n";    // before the @
echo substr($email, $at + 1), "\n";   // after the @

// strstr does the same in one call
echo strstr($email, '@', true), "\n";

Check that strpos() did not return false before using its result: if the character is missing, false is treated as 0 with no warning, so substr('nobody', 0, $at) returns "" and substr('nobody', $at + 1) returns "obody".