[Resolved] yet another regexp question...real simple...

dbmasters

I am trying to make a little function to remove stop words from a search query, I will make the array small for the sake of this example:

Function stopWords($input)
{
$stops=array("after","also","an","and");
$noStops=preg_replace("/".$stops."/i","",$input);
$output=str_replace(" "," ",$noStops);
return stripslashes($output);
}

The $noStops preg_replace I need the RegExp that will check for either a space of beginning/end of string. So it doesn't for example take "an" out of "animal" but only if it is a standalone word.

And please, explain WHY it works briefly as I am trying to understand this RegExp thing...I am sure it is very simple, but I learn by example much better than reading theory.

Thanks guys and gals.

dbmasters

wait, I did read about a \b function so I tried this:

$stops=array("after","also","an","and");
$noStops=preg_replace("\b".$stops."\b","",$input);

but it returned nothing...am I on the right track?

mrhappiness

function removestopwords($input) {
  $stops = array('after', 'also', 'an', 'and');
  $pattern = '% ('.preg_quote(implode('°',$stops), '%').') %i';
echo 'Pattern that can be used almost: '.$pattern;
echo '<br />';
  $pattern = str_replace('°', '|', $pattern);
echo 'Pattern that can be used: '.$pattern;
echo '<br />';
  $nostops = preg_replace($pattern, ' ', $input);
  return preg_replace('%[\\s]+%', ' ', $nostops);
}

echo removestopwords('assa        and fcsd           also  sdvfedfv');

should do the job

have a look ath the echos 🙂

BuzzLY

You can pass an array of patterns directly to preg_replace:

$stops=array("/after/i","/also/i","/an/i","/and/i");
$noStops=preg_replace($stops,"",$input);
$output=preg_replace("/\\s+/"," ",$noStops);
return stripslashes($output);

dbmasters

OK, thanks for the reply...

that returned "a a fc d dvfedfv"

which really took out good characters so that doesn't quite do it...

Also, what is °?

mrhappiness

Originally posted by BuzzLY
You can pass an array of patterns directly to preg_replace

which of course is more elegant than my attempt 🆒

dbmasters

Originally posted by BuzzLY
You can pass an array of patterns directly to preg_replace:
$stops=array("/after/i","/also/i","/an/i","/and/i");
$noStops=preg_replace($stops,"",$input);
$output=preg_replace("/\\s+/"," ",$noStops);
return stripslashes($output);
[/B]

Yes, but won't /an/i remove "an" from "dan"? If I am mistaken here, clue me in 🙂

BuzzLY

Hmmm.... How about this:

$stops=array("after","also","an","and");
foreach($stops as $key=>$value) {
    $stops[$key] = "/\\b$value\\b/i";
}
$noStops=preg_replace($stops,"",$input);
$output=preg_replace("/\\s+/"," ",$noStops);
return stripslashes($output);

I haven't actually tested this yet, but in theory it should work.

mrhappiness

Originally posted by dbmasters
which really took out good characters so that doesn't quite do it...

corrected it

the forum didn't display my backslash in [\s]

i just fixed it, try again please

Also, what is °?

to seperate serveral words that match you use the | (match1|match2) but the | get's escaped by preg_quote so that it is | which is not useful at all

therefor i use ° instead of | and afterwards replace it with |

assuming that the charcater ° won't appear on the list...

dbmasters

PERFECT!

mrhappiness

i live to please

and don't forget to mark this thread as resolved (if it's resolved now)

dbmasters

Originally posted by BuzzLY
Hmmm.... How about this:
$stops=array("after","also","an","and");
foreach($stops as $key=>$value) {
    $stops[$key] = "/\\b$value\\b/i";
}
$noStops=preg_replace($stops,"",$input);
$output=preg_replace("/\s+/"," ",$noStops);
return stripslashes($output);
I haven't actually tested this yet, but in theory it should work. [/B]

This one worked perfectly...so, OK, I think I get why that worked too...cool, thanks folks, phpbuilder always comes thru in the clutch. 🙂

BuzzLY

The great thing is that because you are using word boundaries, it will still remove the stop words if they are in quotes, or next to punctuation.