preg_match_all all combinations with word bounderies - php

Question

i've got the following string:
$string = "König Friedrich August III. von Sachsen - Adel Sachsen, Waidmannsheil, Kapitaler 16ender erlegt auf der Jagd am 2. Oktober 1905, gelaufen 30.06.1909, Verlag, Karlowa Walter, Dresden";
Now I wan't to find words in that string using preg_match_all:
preg_match_all("/\b(abituria)\b|\b(absolvia)\b|\b(adel sachsen)\b|\b(adel)\b|\b(sachsen)\b|\b(könig)\b/i",$string,$matches);
The string matches only for
array(
0 => "König",
1 => "Adel Sachsen"
)
but I need that it also returns "Adel" in the $matches-Array.
How can I do that? I think my problem is that: "After the first match is found, the subsequent searches are continued on from end of the last match."
Update
That does not work:
preg_match_all('/(?=\b(adel sachsen|adel)\b)/ui', $string, $matches);
print_r($matches[1]);
Array
(
[0] => Adel Sachsen
)
preg_match_all('/(?=\b(adel|adel sachsen)\b)/ui', $string, $matches);
print_r($matches[1]);
Array
(
[0] => Adel
)
But i need the following as result:
Array
(
[0] => Adel Sachsen,
[1] => Adel
)

Answer 1

I would just search for each word/combination (generate a pattern for each) and map the according match to the result array or set false, if it doesn't match. Then filter the false elements:
$arr = ["nadel", "adel", "knödel", "sachsen", "adel sachsen"];
$str = "Friedrich August III. von Sachsen - Adel Sachsen";
$res = array_filter(array_map(function ($s) use (&$str) {
$s = '/\b'.preg_quote($s,'/').'\b/iu';
return preg_match($s, $str, $out) ? $out[0] : false; }, $arr));
sort($res); print_r($res);
See test at eval.in (anonymous functions with array_map: at least PHP 5.3 is required)
Array
(
[0] => Adel
[1] => Adel Sachsen
[2] => Sachsen
)
The function can be further improved to return arrays, if such as different cases for same words is desired or capturing the offset.

Answer 2

You can use lookahead to get your ovelaping matches:
preg_match_all('/(?=\b(abituria|absolvia|adel sachsen|adel|sachsen|könig)\b)/ui',
$string, $matches);
print_r($matches[1]);
Array
(
[0] => König
[1] => Sachsen
[2] => Adel Sachsen
[3] => Sachsen
)
RegEx Demo
Update: Based on your updated code snippet you can do this:
preg_match_all('/(?=\b(adel sachsen)\b)(?=\b(adel)\b)/ui', $string, $matches);
unset($matches[0]);
print_r($matches);
Output:
Array
(
[1] => Array
(
[0] => Adel Sachsen
)
[2] => Array
(
[0] => Adel
)
)

Answer 3

As you already noticed, preg_match_all continues searching after the end of each last match, so it is not the best tool for your task.
The easy but less performant solution would be to do one preg_match for each single search term instead.
If the strings are not much longer than your example I would go for this, optimizing it seems not to be worth it.
If performance is really critical, I would group prefixes of other terms with them, ordering each group by longest term first:
abituria
absolvia
adel sachsen, adel
sachsen
könig
Now use the regex with lookahead assertion:
preg_match_all('/(?=\b(abituria|absolvia|adel sachsen|adel|sachsen|könig)\b)/ui',
$string, $matches);
If $string contains "adel", but not "adel sachsen", it will match correctly. If it contains "adel sachsen", it will only match "adel sachsen", but from the groups that we constructed before, we know that it also matches prefixes of "adel sachsen", i.e. "adel".