PHP preg_match(): Regular Expressions with Examples

preg_match($pattern, $subject, $matches) checks whether a string matches a regular expression. It returns 1 on a match and 0 otherwise, and fills $matches with what it found: preg_match('/\d+/', 'Order 4521 shipped', $m) returns 1 and sets $m[0] to "4521".

<?php
$text = 'Order 4521 shipped on 2026-03-14';

if (preg_match('/\d+/', $text, $m)) {
    echo "First number: {$m[0]}\n";
}

var_dump(preg_match('/shipped/', $text));   // found
var_dump(preg_match('/cancelled/', $text)); // not found
var_dump(preg_match('/SHIPPED/i', $text));  // i = case-insensitive

The pattern is a string with delimiters around it, usually /.../, followed by modifiers such as i. Write patterns in single quotes: inside double quotes PHP first expands $name and escapes such as \x41 or \101, so the regex engine can receive a different pattern from the one you typed.

Capture groups

Parentheses in a pattern capture the part of the string they match. $m[0] is the whole match, $m[1] the first group, $m[2] the second, counted by opening parenthesis from the left.

<?php
$text = 'Shipped on 2026-03-14 at 09:30';

if (preg_match('/(\d{4})-(\d{2})-(\d{2})/', $text, $m)) {
    print_r($m);
    [, $year, $month, $day] = $m;
    echo "Day $day, month $month, year $year\n";
}

// (?: ) groups without capturing
preg_match('/(?:at|on) (\d{2}:\d{2})/', $text, $t);
print_r($t);

Named groups

(?<name>...) gives a group a name. The value is stored under the name and under its number, so code that reads the result does not break when someone adds a group earlier in the pattern.

<?php
$log = '[2026-03-14 09:30:12] ERROR: Disk almost full';

$pattern = '/^\[(?<date>[\d-]+) (?<time>[\d:]+)\] (?<level>\w+): (?<message>.+)$/';

if (preg_match($pattern, $log, $m)) {
    echo $m['level'], "\n";
    echo $m['message'], "\n";
    echo $m['date'], ' ', $m['time'], "\n";
}

// Keep only the named keys
print_r(array_filter($m, 'is_string', ARRAY_FILTER_USE_KEY));

Find every match with preg_match_all

preg_match_all() keeps searching after the first match and returns the number of matches. By default $m[0] holds all full matches and $m[1] all values of the first group. With PREG_SET_ORDER you get one array per match instead, which is easier to loop over.

<?php
$text = 'Contact ana@example.com or bo@test.org, not support@';

$count = preg_match_all('/(\w+)@(\w+\.\w+)/', $text, $m);
echo "Found $count addresses\n";
print_r($m[0]);   // full matches
print_r($m[2]);   // domains

preg_match_all('/(\w+)@(\w+\.\w+)/', $text, $sets, PREG_SET_ORDER);
foreach ($sets as $s) {
    echo "user {$s[1]} at {$s[2]}\n";
}

// Every number in a string
preg_match_all('/\d+/', 'a1 b22 c333', $nums);
print_r($nums[0]);

Validate input: anchors, phone numbers, postal codes

For validation the pattern must match the whole string, so start it with ^ and end it with $. One trap: $ also matches before a final newline, so '/^\d+$/' accepts "123\n". Add the D modifier, or use \z, which only matches at the very end.

<?php
$tests = ['123', '12a', "123\n", ''];

foreach ($tests as $t) {
    printf("%-8s plain:%d  with D:%d\n",
        json_encode($t),
        preg_match('/^\d+$/', $t),
        preg_match('/^\d+$/D', $t)
    );
}

// Japanese postal code: 123-4567
var_dump(preg_match('/^\d{3}-\d{4}$/D', '150-0002'));

// US phone: (555) 123-4567 or 555-123-4567 or 5551234567
$phone = '/^(\(\d{3}\) ?|\d{3}-?)\d{3}-?\d{4}$/D';
foreach (['(555) 123-4567', '555-123-4567', '5551234567', '55-1234'] as $p) {
    echo $p, ' => ', preg_match($phone, $p), "\n";
}

For email addresses, a regex either rejects valid addresses or accepts invalid ones. Use filter_var($email, FILTER_VALIDATE_EMAIL) instead (see filter_var()).

The u modifier for UTF-8 and Japanese text

Without the u modifier, a pattern works on bytes. A Japanese character is three bytes in UTF-8, so . matches a third of a character and lengths come out wrong. With u, the pattern works on characters and you can use Unicode property classes.

<?php
$word = '日本語';

var_dump(preg_match('/^.{3}$/', $word));   // bytes: 9, no match
var_dump(preg_match('/^.{3}$/u', $word));  // characters: 3, match

$text = 'PHPで正規表現を使う';
preg_match_all('/\p{Han}+/u', $text, $kanji);
preg_match_all('/\p{Katakana}+/u', 'コーヒーとカフェラテ', $kata);
print_r($kanji[0]);
print_r($kata[0]);

// Only hiragana (and the long vowel mark)?
var_dump(preg_match('/^[\p{Hiragana}ー]+$/uD', 'ひらがな'));

u also validates the subject: if the string is not valid UTF-8, preg_match() returns false instead of matching.

Escape user input with preg_quote

Characters like ., +, ?, ( and / have special meanings in a pattern. If part of the pattern comes from a variable, such as a search term, pass it through preg_quote() with the delimiter as the second argument, or 1+1 will match 11 and an unescaped / breaks the pattern.

<?php
$search = '1+1';
$text = 'Is 11 the answer? No, 1+1 is 2.';

// Wrong: + means "one or more"
preg_match('/' . $search . '/', $text, $bad);
echo "Unescaped matched: {$bad[0]}\n";

// Right
$pattern = '/' . preg_quote($search, '/') . '/';
echo "Pattern: $pattern\n";
preg_match($pattern, $text, $good);
echo "Escaped matched: {$good[0]}\n";

An invalid pattern does not throw. PHP prints a warning and preg_match() returns false:

Warning: preg_match(): No ending delimiter '/' found in /path/to/index.php on line 3

So check === false (or preg_last_error_msg()) when the pattern is built at runtime. For a plain substring check with no pattern at all, str_contains() is faster and clearer. To change the text you matched, use preg_replace().