Question

I need detect words that separate with space in a text. for example my text is:

some parent +kid -control "human right" world

now I need to detect some, parent, world. (all words that don’t have + – ( ) < > before and after, and all words inside quotes must be discarded) so I write this regex with preg_match_all():

(?:^|[\s]+)((?:(?![\+\(\)\<\>\s\-\"]).)+)(?:[\s]+|$)

but it only detect some and world. how can I fix it?

EDIT

I need it for Javascript too. But it seems it doesn’t work for Javascript. how can I do it with javascript?

EDIT

I found a solution but it seems stupid way. what is your ideas?

$str = 'some parent +kid -control "my human right" world';
$words=array();
$quot=false;
$discard=false;
$word='';
for($i=0;$i<=strlen($str);$i++){
    $chr=substr($str,$i,1);
    if($chr=='"'){
        if($quot){
            $quot=false;
        }else{
            $quot=true;
        }
        continue;
    }
    if($quot)continue;
    if($chr==' '||$i==strlen($str)){
        if(strlen($word)&&!$discard)$words[]=$word;
        $discard=false;
        $word='';
        continue;
    }elseif(in_array($chr,array('+','-','(',')','<','>'))){
        $discard=true;
        continue;
    }
    $word.=$chr;
}
print_r($words);//Array ( [0] => some [1] => parent [2] => world ) 

EDIT
Final way for PHP (this is for multi-language queries) (special thanks to rubber boots):

$query='some parent +kid -control "my human right" world';
$result=array();
if(preg_match_all('/(?:"[^"]+")|(?:^|[\s])(?P<q>(?:(?![\+\(\)\<\>\s\-\"]).)+)/',$query,$match)){
    $result=array_filter($match['q'],'strlen');
}
print_r($result);// some,parent,world

Final way for javascript (this is for multi-language queries) (special thanks to rubber boots):

var query='some parent +kid -control "my human right" world';
var result=Array();
var tmp;
var patt=RegExp('(?:"[^"]+")|(?:(?:^|\\s)((?:(?![\\+\\(\\)\\<\\>\\s\\-\\"]).)+))', 'g');
while(tmp = patt.exec(query)){
    if(typeof(tmp[1])!=='undefined') result.push(tmp[1]);
}
alert(result);// some,parent,world

Answer 1

If the following string is given:

 $t ='some parent +kid -control "human huhu right" world';

it’s possible to extract words according to your specification with a rather simple expression too:

 $r = '/ (?:" [^"]+ ")? \s?
         (?<!\S) \b (\w+)
       /x';
 preg_match_all($r, $t, $matches);

This results in:

foreach($matches[1] as $m) echo $m . "\n";

some
parent
world

The technique used:

The expr (?:" [^"]+ ")? consumes the quotes and their contents.

---

Addendum: Javascript

For Javascript, you need to use a slightly more complicated approach, Javascript has no lookbehind assertions, we fake them with (?:^|\\s) in front of an allowed word.

This will work:

  var t = 'some parent +kid -control "human huhu right" world';
  var r = /(?:"[^"]+")?(?:^|\s)(\b\w+)/g;
  var a = [];
  while(m = r.exec(t)) a.push(m[1]);

We use the same technique here – generate captured submatches in $1 for the words we need.

The contents of the array a, (document.getElementById("myhtml").innerHTML = a;) will contain then:

some,parent,world