Regular Expressions (regexp/PCRE)

Syntax#

• preg_replace($pattern, $replacement, $subject, $limit = -1, $count = 0);
• preg_replace_callback($pattern, $callback, $subject, $limit = -1, $count = 0);
• preg_match($pattern, $subject, &$matches, $flags = 0, $offset = 0);
• preg_match_all($pattern, $subject, &$matches, $flags = PREG_PATTERN_ORDER, $offset = 0);
• preg_split($pattern, $subject, $limit = -1, $flags = 0)

Parameters#

All PCRE strings in PHP must be enclosed with delimiters. A delimiter can be any non-alphanumeric, non-backslash, non-whitespace character. Popular delimiters are ~, /, % for instance.

PCRE patterns can contain groups, character classes, character groups, look-ahead/look-behind assertions and escaped characters.

It is possible to use PCRE modifiers in the $pattern string. Some common ones are i (case insensitive), m (multiline) and s (the dot metacharacter includes newlines). The g (global) modifier is not allowed, you will use the preg_match_all function instead.

Matches to PCRE strings are done with $ prefixed numbered strings:

<?php

$replaced = preg_replace('%hello ([a-z]+) world%', 'goodbye $1 world', 'hello awesome world');

echo $replaced; // 'goodbye awesome world'

String matching with regular expressions

preg_match checks whether a string matches the regular expression.

$string = 'This is a string which contains numbers: 12345';

$isMatched = preg_match('%^[a-zA-Z]+: [0-9]+$%', $string);
var_dump($isMatched); // bool(true)

If you pass in a third parameter, it will be populated with the matching data of the regular expression:

preg_match('%^([a-zA-Z]+): ([0-9]+)$%', 'This is a string which contains numbers: 12345', $matches);
// $matches now contains results of the regular expression matches in an array.
echo json_encode($matches); // ["numbers: 12345", "numbers", "12345"]

$matches contains an array of the whole match then substrings in the regular expression bounded by parentheses, in the order of open parenthesis’s offset. That means, if you have /z(a(b))/ as the regular expression, index 0 contains the whole substring zab, index 1 contains the substring bounded by the outer parentheses ab and index 2 contains the inner parentheses b.

Split string into array by a regular expression

$string = "0| PHP 1| CSS 2| HTML 3| AJAX 4| JSON";

//[0-9]: Any single character in the range 0 to 9
// +   : One or more of 0 to 9
$array = preg_split("/[0-9]+\|/", $string, -1, PREG_SPLIT_NO_EMPTY);
//Or
// []  : Character class
// \d  : Any digit
//  +  : One or more of Any digit
$array = preg_split("/[\d]+\|/", $string, -1, PREG_SPLIT_NO_EMPTY);

Output:

Array
(
    [0] =>  PHP
    [1] =>  CSS 
    [2] =>  HTML 
    [3] =>  AJAX 
    [4] =>  JSON
)

To split a string into a array simply pass the string and a regexp for preg_split(); to match and search, adding a third parameter (limit) allows you to set the number of “matches” to perform, the remaining string will be added to the end of the array.

The fourth parameter is (flags) here we use the PREG_SPLIT_NO_EMPTY which prevents our array from containing any empty keys / values.

String replacing with regular expression

$string = "a;b;c\nd;e;f";
// $1, $2 and $3 represent the first, second and third capturing groups
echo preg_replace("(^([^;]+);([^;]+);([^;]+)$)m", "$3;$2;$1", $string);

Outputs

c;b;a
f;e;d

Searches for everything between semicolons and reverses the order.

Global RegExp match

A global RegExp match can be performed using preg_match_all. preg_match_all returns all matching results in the subject string (in contrast to preg_match, which only returns the first one).

The preg_match_all function returns the number of matches. Third parameter $matches will contain matches in format controlled by flags that can be given in fourth parameter.

If given an array, $matches will contain array in similar format you’d get with preg_match, except that preg_match stops at first match, where preg_match_all iterates over the string until the string is wholly consumed and returns result of each iteration in a multidimensional array, which format can be controlled by the flag in fourth argument.

The fourth argument, $flags, controls structure of $matches array. Default mode is PREG_PATTERN_ORDER and possible flags are PREG_SET_ORDER and PREG_PATTERN_ORDER.

Following code demonstrates usage of preg_match_all:

$subject = "a1b c2d3e f4g";
$pattern = '/[a-z]([0-9])[a-z]/';

var_dump(preg_match_all($pattern, $subject, $matches, PREG_SET_ORDER)); // int(3)
var_dump($matches);
preg_match_all($pattern, $subject, $matches); // the flag is PREG_PATTERN_ORDER by default
var_dump($matches);
// And for reference, same regexp run through preg_match()
preg_match($pattern, $subject, $matches);
var_dump($matches);

The first var_dump from PREG_SET_ORDER gives this output:

array(3) {
  [0]=>
  array(2) {
    [0]=>
    string(3) "a1b"
    [1]=>
    string(1) "1"
  }
  [1]=>
  array(2) {
    [0]=>
    string(3) "c2d"
    [1]=>
    string(1) "2"
  }
  [2]=>
  array(2) {
    [0]=>
    string(3) "f4g"
    [1]=>
    string(1) "4"
  }
}

$matches has three nested arrays. Each array represents one match, which has the same format as the return result of preg_match.

The second var_dump (PREG_PATTERN_ORDER) gives this output:

array(2) {
  [0]=>
  array(3) {
    [0]=>
    string(3) "a1b"
    [1]=>
    string(3) "c2d"
    [2]=>
    string(3) "f4g"
  }
  [1]=>
  array(3) {
    [0]=>
    string(1) "1"
    [1]=>
    string(1) "2"
    [2]=>
    string(1) "4"
  }
}

When the same regexp is run through preg_match, following array is returned:

array(2) {
  [0] =>
  string(3) "a1b"
  [1] =>
  string(1) "1"
}

String replace with callback

preg_replace_callback works by sending every matched capturing group to the defined callback and replaces it with the return value of the callback. This allows us to replace strings based on any kind of logic.

$subject = "He said 123abc, I said 456efg, then she said 789hij";
$regex = "/\b(\d+)\w+/";

// This function replaces the matched entries conditionally 
// depending upon the first character of the capturing group
function regex_replace($matches){
    switch($matches[1][0]){
        case '7':
            $replacement = "<b>{$matches[0]}</b>";
            break;
        default:
            $replacement = "<i>{$matches[0]}</i>";
    }
    return $replacement;
}

$replaced_str = preg_replace_callback($regex, "regex_replace", $subject);

print_r($replaced_str);
# He said <i>123abc</i>, I said <i>456efg</i>, then she said <b>789hij</b>