term highlight algorithm (HTML)

Question

Hi, I know there are quite some term highlighting questions asked but as far as I know none answers mine. The search terms are put into an array

$keyarray = array("DE", "ABCD");
$string = "ABCDEF";
foreach ($keyarray as $value) {
   $string = str_ireplace($value, "<b>{$value}</b>", $string);
}

The results will obviously be ABCDEF rather than ABCDEF So its there anyway way that I can highlight both terms using a BOLD tag extremely fast using PHP?

Answer 1

maybe use regexp like: /de|acbd/. regex is searching from left to right and it's not going back so there will be no situtation like <b>1111<b></b>112222</b>

Answer 2

You may try this:

$keyarray = array("DE", "ABCD");
$string = "ABCDEF";
$words = explode(' ',$string);
foreach($words as &$w){
   foreach($keyarray as $criteria) {
        if(stripos($w,$criteria)!==false){
            $w = "<b>$w</b>";
                    break;//avoid duplicate wrapping
        }
   }
}
$string = implode(' ',$words);

Answer 3

"Extremely fast" is a relative term. That aside, you have a couple of options:

• Regular Expressions: if you are very good with regexes, this is a valid use for them. You can also look forward/behind with them, which allows for a fair amount of flexibility.

http://www.regular-expressions.info/lookaround.html

• Character by character parse: this is often the best way of doing string manipulation (and fastest), but can be the most time-consuming to create.

• String replaces; fast but there's usually an edge case that doesn't work correctly (speaking from experience).

In all of these scenarios, you can benefit from pre-optimizing your list of terms by sorting/grouping/filtering them appropriately. For example, sorting biggest to smallest length would ensure that you didn't split up a long string (and miss a match) by bolding a shorter string within it.

You could also predetermine optimal regex(es) by examining all the search terms before beginning the replace. Again, this would assume you are pretty savvy with regexes.