regEx: only full words - only once...

sid

ok - my problem is that i am working on a website for a big ski company. they have descriptions for their ski models. then there is a database table that includes "technology" terms. if one of these terms is found in a description it should be turned into a link to a description for it.

ok - my problem is now: there is one technolgy calles "as" for example: how do i force the code to only replace full words? (and not the "as" in "mASsses") - also, if i already replaced one string: how do i not re-replace (flag) it (f.ex.: 2 terms: "diagonal" and "diagonal force" - diagonal force is first in the database - so it eventually looks like this:

<a href="link.php?id=diagonalforce"><a href="link.php?id=diagonal">diagonal</a> force</a>

since diagonal gets parsed twice.)

ok - my current code looks like this:

function techReplace($text, $cat = 1)
{
//FILL ARRAY: $techResult[ID] = TITLE;

$sql = "SELECT id, title FROM products_technik WHERE cat = " . $cat;
$techResult = mysql_query($sql);
while ($techRow = mysql_fetch_array($techResult)) {
	$techKey = $techRow[title];
	$technology[$techKey] = "<a href=\"link.php?id=" . $techRow[id] . "\">" . ucwords(strtolower($techRow[title])) . "</A>";
}

$textReturn = $text;

//PARSE TEXT
	if (count($technology)) {
		foreach ($technology as $key => $value) {
			$textReturn = eregi_replace($key, $value, $textReturn);
		}
	}

return $textReturn;
}

thanks a lot in advance!!!

Weedpacket

If you use preg_match() instead of eregi_match(), you'll be able to use \b to indicate word boundaries.

Eg., "\bas\b" matches " as " but not "mass " or " ascent" or " bassoon ".

sid

thanks a lot man...

would you mind posting the right preg_replace() version???

preg_replace kinda is not my best friend...

Weedpacket

I almost did:

/\bas\b/i

Needless to say, you'd replace "as" with whatever it is you're really looking for; the "i" is to make the search case-insensitive. Note also that the preg_replace() function can take arrays of keys and values in a single call, so you wouldn't need to call it in a loop. (Of course, you'd probably need a loop to build the arrays. Swings and roundabouts I guess...if you can combine the array-building loop with the loop before it, that would be a saving.)

sid

hi weedpacket...

first of all thanks for the help. i know you gave me a big clue already. but i just cant seem to manage to make this work. maybe im too retarded. i know i can have this working with arrays. so how would the preg_replace() line look like then?

say i have 2 (already filled) arrays - the keys of each array correspond:

$technologyReplace[] = "<a href=\"link.php?id=" . $techRow[id] . "\">" . ucwords(strtolower($techRow[title])) . "</A>"; 
$technologyPattern[] = $techRow[title];

would i do it like that then???

$textReturn = preg_replace("/\b" . $technologyPattern[]  . "\b/i", $technologyReplace[], $rext);

???

as far as i remember i got a delimiter error last time i tired it... any ideas where i got the twist in my system???

thanks a lot dude!
rock on - sid
xxx