PHP: How to split a string by dash and everything between brackets. (preg_split or preg_match)

Question

I've been wrapping my head around this for days now, but nothing seems to give the desired result.

Example:

$var = "Some Words - Other Words (More Words) Dash-Binded-Word";

Desired result:

array(
[0] => Some Words
[1] => Other Words
[2] => More Words
[3] => Dash-Bound-Word
)

I was able to get this all working using preg_match_all, but then the "Dash-Bound-Word" was broken up as well. Trying to match it with surrounding spaces didn't work as it would break all the words except the dash bound ones.

The preg_match_all statement I used (which broke up the dash bound words too) is this:

preg_match_all('#\(.*?\)|\[.*?\]|[^?!\-|\(|\[]+#', $var, $array);

I'm certainly no expert on preg_match, preg_split so any help here would be greatly appreciated.

Answer 1

You can use a simple preg_match_all:

\w+(?:[- ]\w+)*

See demo

• \w+ - 1 or more alphanumeric or underscore
• (?:[- ]\w+)* - 0 or more sequences of...
  • [- ] - a hyphen or space (you may change space to \s to match any whitespace)
  • \w+ - 1 or more alphanumeric or underscore

IDEONE demo:

$re = '/\w+(?:[- ]\w+)*/'; 
$str = "Some Words - Other Words (More Words) Dash-Binded-Word"; 
preg_match_all($re, $str, $matches);
print_r($matches[0]);

Result:

Array
(
    [0] => Some Words
    [1] => Other Words
    [2] => More Words
    [3] => Dash-Binded-Word
)

Answer 2

You can split by:

/\s*(?<!\w(?=.\w))[\-[\]()]\s*/

Explanation:

1. The match is attempted against the character class [\-[\]()] (matches any of those characters). You could also add any char you want to that character class.
2. It's using a negative lookbehind (?<!\w) for the condition: "not preceded by a word character".
3. And it also has a nested lookahead (?=.\w) that checks for: "if the first condition is met, it shouldn't be followed by any char -the one used to split- and a word character".
4. \s* at the beggining and the end is to trim whitespaces.

Code:

$input_line = "Some Words - Other Words (More Words) Dash-Binded-Word";
$result = preg_split("/\s*(?<!\w(?=.\w))[\-[\]()]\s*/", $input_line);
var_dump($result);

Output:

array(4) {
  [0]=>
  string(10) "Some Words"
  [1]=>
  string(11) "Other Words"
  [2]=>
  string(10) "More Words"
  [3]=>
  string(16) "Dash-Binded-Word"
}

Run this code here

Capturing parens

As stated in another comment, if you want to also capture parentheses:

$result = preg_split("/\s*(?:(?<!\w)-(?!\w)|(\(.*?\)|\[.*?]))\s*/", $input_line, -1, PREG_SPLIT_DELIM_CAPTURE);

Answer 3

Modifying the input string to suit any particular exploding technique would be indirect and indicate that a suboptimal exploding technique is being used.

The truth is, your required logic can be boiled down to: "explode on each sequence of non-word characters that have a length of 2 or more". This is what that pattern looks like with preg_split().

Code: (Demo)

$var = "Some Words - Other Words (More Words) Dash-Binded-Word";

var_export(preg_split('~\W{2,}~', $var));

Output:

array (
  0 => 'Some Words',
  1 => 'Other Words',
  2 => 'More Words',
  3 => 'Dash-Binded-Word',
)

It doesn't get any simpler than that.

Answer 4

Try this (combination of str_replace and explode). It is not optimum but may work for this case:

$var = "Some Words - Other Words (More Words) Dash-Binded-Word";
$arr = Array(" - ", " (", ") ");
$var2 = str_replace($arr, "|", $var);
$final = explode('|', $var2);
var_dump($final);

Output:

array(4) { [0]=> string(10) "Some Words" [1]=> string(11) "Other Words" [2]=> string(10) "More Words" [3]=> string(16) "Dash-Binded-Word" }

Answer 5

$var = "Some Words - Other Words (More Words) Dash-Binded-Word";

$var=preg_replace('/[^A-Za-z\-]/', ' ', $var);
$var=str_replace('-', ' ', $var); // Replaces all hyphens with spaces.
print_r (explode(" ",preg_replace('!\s+!', ' ', $var)));  //replaces all multiple spaces with one and explode creates array split where there is space

OUTPUT :-

Array ( [0] => Some [1] => Words [2] => Other [3] => Words [4] => More [5] => Words [6] => Dash [7] => Binded [8] => Word )