PHP
Capture groups
Learn to create capture groups with parentheses () and use backreferences.
Capture groups: (...)
Wrapping a piece of the pattern in parentheses makes it a capture group. After preg_match, groups are numbered as $matches[1], $matches[2].
For example, the pattern /(\d{2})\.(\d{2})/ on "29.05" gives $matches[1] = "29" and $matches[2] = "05". $matches[0] always holds the full match.
In preg_replace, captured groups are referenced as $1, $2. To swap date parts, preg_replace("/(\d{2})\.(\d{2})/", "$2.$1", "29.05") works.
Named groups: (?P<name>...)
Naming groups instead of using numbers makes code readable. Write (?P<name>\w+); the match is read with $matches['name'].
Example: after preg_match("/(?P<year>\d{4})/", "in 2026", $m), $m['year'] holds "2026". The number still works: $m[1] is the same.
Since PHP 7, (?<name>...) and (?P<name>...) are equivalent. Named groups are preferred for readability.
Exercise Split first and last name
In $line = "Ayşe Demir";, use preg_match with two capture groups to split first and last name. Write the pattern into $pattern as "/(\w+) (\w+)/u"; the u modifier is needed for the Turkish letters. Print $matches[1] and $matches[2] on separate lines.
Show example solution
Set pattern to "/(\w+) (\w+)/u". Call preg_match($pattern, $line, $matches). Print with echo $matches[1] . "\n"; and echo $matches[2] . "\n";
Example solution:
<?php
// Split first and last name with preg_match capture groups
$line = "Ayşe Demir";
// Set the pattern and matches
$pattern = "/(\w+) (\w+)/u";
preg_match($pattern, $line, $matches);
// Print first and last name
echo $matches[1] . "\n";
echo $matches[2] . "\n";
Test yourself
-
In preg_replace, how do you reference the whole match?
Show answer
$0 — Correct. $0 is the whole match; $1 and onward are the groups.
-
Group syntaxes
- (\w+)
- Unnamed capture group
- (?P<name>\w+)
- Named capture group
- $1
- Reference to the first group
- (?:\w+)
- Non-capturing group
-
Swap the groups
$new = preg_replace("/(\w+) (\w+)/", "$____ $1", "Ali Veli");
Show answer
$new = preg_replace("/(\w+) (\w+)/", "$2 $1", "Ali Veli");
Next lesson: Summary - Regex
