How to split string on multiple delimiters keeping some delimiters?

Question

i'm looking for a way to split a string in several words, based on some delimiters.

For example the string word1&word2 !word3 word4 &word5 should give me an array with the following words:

word1
&word2
!word3
word4
&word5

How to reach that? I tried several solution with str_replace() but i cannot figure out the best way to obtain what i need. Maybe the solution could be using regular expressions, but i do not know how to use them.

you should look how to use explode, here http://php.net/explode — keikoku92, Jun 10 '15 at 15:41
Obviously i have already seen that page but there isn't anything could help me. Am i wrong? — smartmouse, Jun 10 '15 at 15:42
@keikoku92 `explode` will only work when there's a single consistent delimiter in the string like a comma. Here there are multiple possibilities. — Ankh, Jun 10 '15 at 15:43
you need to decide to one delimiter (for explode) or one rule. in your example this would be after each numeric character, is that what you mean? — Michael B., Jun 10 '15 at 15:44
as a board idea, you can 1) replace useless characters as &! or even the space character by a delimiter of your choice in your string and then explode it as it will have teh format needed — keikoku92, Jun 10 '15 at 15:44
You can just split on spaces, or one of the non-word characters per lookahead `\s+|(?=\W)`. You need to be more specific on what constitutes a delimiter. — mario, Jun 10 '15 at 15:48
I think my example is self-explaining. It explains better that i could do. — smartmouse, Jun 10 '15 at 15:50
You could use preg_split, as in [this](http://stackoverflow.com/questions/1452777/how-to-split-a-string-by-multiple-delimiters-in-php) answer. — fkorsa, Jun 10 '15 at 15:51
You think bad, your example isn't clear at all and can be subject to many interpretations. — Casimir et Hippolyte, Jun 10 '15 at 15:52
I have already seen that answer but as you can see i need to keep the delimiters attached to the splitted words. — smartmouse, Jun 10 '15 at 15:53

Carsten Massmann · Accepted Answer · 2015-06-10T16:16:20.793

2

Try this:

$src='word1&word2 !word3 word4 &word5';
$arr=explode(' ',$src=preg_replace('/(?<=[\w])([&!])/',' $1',$src));
echo join('<br>',$arr); // present the result ...

First change any occurence of a group consisting of a single character of class [&!] that is preceded by a 'word'-character into ' $1' (=itself, preceded with a blank) and then explode()the string using the blanks as separators.

If you need to deal with multiple blanks as separators between the words you could of course replace the (faster) explode(' ',...) with a slighty more "refined" preg_split('/ +/',...).

edited Jun 10 '15 at 16:16

answered Jun 10 '15 at 15:54

Carsten Massmann

26,510
2
22
43

It works, but how to edit it if i encounter strings that contains `word1!word2`. In this case it doesn't split. – smartmouse Jun 10 '15 at 16:00
Just edited my reply: use a group consisting of a single character belonging to the character class `[&!]` instead of the single character `&` and replace it with its own occurence (`$1`) . – Carsten Massmann Jun 10 '15 at 16:04
The reason for what i asked this it's because i'm trying to parse users search strings. Maybe you can help me with this: http://stackoverflow.com/questions/30760625/how-to-parse-user-search-string-for-postgresql-query – smartmouse Jun 10 '15 at 16:21

score 1 · Answer 2 · answered Jun 10 '15 at 16:07

You can do use of preg_split, and do some customization as needed. see example below:-

function customExplode($string){
    if($matches = preg_split('/[\s&!]+/i', $string, null, PREG_SPLIT_OFFSET_CAPTURE)){
        $return = array();
        foreach ($matches as $match) {
            $return[] = (($match[1]-1) >= 0) ? substr($string, $match[1]-1, 1).$match[0] : $match[0];
        }
        return $return;
    } else {
        return $string;
    }
}
$word = 'word1&word2 !word3 word4 &word5';
print_r(customExplode($word));

Results

Array
(
    [0] => word1
    [1] => &word2
    [2] => !word3
    [3] =>  word4
    [4] => &word5
)

score 0 · Answer 3 · answered Mar 12 '22 at 10:28

Previous answers have overcomplicated this task.

The required coding/pattern logic is to split on each space OR on the position before & or !.

Code #1: (Demo) - split on space or position between non-space and a & or !

var_export(
    preg_split('/ |(?<! )(?=[&!])/', $string)
);

Code #2: (Demo) - hastily split on spaces or positions before & or ! and omit empty elements from result

var_export(
    preg_split('/ |(?=[&!])/', $string, 0, PREG_SPLIT_NO_EMPTY)
);

Both Output:

array (
  0 => 'word1',
  1 => '&word2',
  2 => '!word3',
  3 => 'word4',
  4 => '&word5',
)

How to split string on multiple delimiters keeping some delimiters?

3 Answers3