pattern matching ?

[Anon]

Anyone know of a regexp or similar function that would find (by ignoring a pattern <somecode>) the word "foo" in the following string:

aaa bbb ccc f<somecode>oo ddd eee

and replace it so the resulting string would be:

aaa bbb ccc <b>f<somecode>oo</b> ddd eee

In addition, <somecode> could be embedded anywhere in "foo", f<somecode>oo OR fo<somecode>o

Thanks, Dan

[Anon]

To clarify, since I didn't escape the html bold codes, the bold f<somecode>oo has obviously be surrounded with html bold and bold-off codes--that was the INTENDED replacement. Thanks!

[Anon]

It seems reasonable to guess that you're wanting to embolden the entire word (and just the word) which <somecode> appears - but you don't mind if <somecode> hangs around afterwards.

I wouldn't want to try this with the ereg functions...

$string preg_replace("/(\b.?<somcecode>.?\b)\b/","<b>\1</b>",$string);

[Anon]

Oops, an extra word-boundary marker slipped in there - drop the last \b in the search string 🙂

And of course, there are bold tags :🙂)

And an = sign might be useful at some point ::🙂))

And of course it's "somecode", not "somcecode" :::🙂)))

Sorry, I've got some 120db thrash metal playing in my ear right now...

[Anon]

Hi Hayley,

Thanks for your input. It got me on the right track. After messing around with different codes, maybe the following will explain where I'm at:

NOTE: Snippet result should surround all words that contain "999" with parenthesis. I used this test to get rid of the bolding confusion.

Thanks...

$tempstr = "aaa bbb ccc ddd eee999eee fff ggg eee999eee sdf";

echo $tempstr . "<br>";

$tempstr = preg_replace("/(\b.?999.?\b)/","(\1)",$tempstr); // the first word boundary marker seems to be greedy

/ ADDITIONAL ATTEMPT
$tempstr = preg_replace("/(.)(\b.?999.?\b)/","\1(\2)",$tempstr); // grabs only last occurance
*/
echo $tempstr;

[Anon]

It's quieter now 🙂

/(\w999\w)/

\w: "wordlike" characters

[Anon]

Thanks again, Hayley, for your input.
I have included the following code to show the success so far, related back to my initial question.

<?
$tempstr = "aaa bbb eeeeee ddd eeeCCCeee fff ggg eCCCeeeee";
$srchword = "eeeeee";
$somecode = "CCC"; // variable, any code, in reality is an html-like code
$boldon = "BB"; // instead of html codes--just for this test
$boldoff = "bb"; // instead of html codes--just for this test

      $count = strlen($srchword)-1;

      echo $tempstr . "<br>";
      $tempstr = preg_replace("/(\w*". preg_replace("/(.)/","\\1[" . $somecode . "]*",$srchword,$count) ."\w*)/i",$boldon . "\\1" . $boldoff,$tempstr);
      echo $tempstr;

?>

Curious about 2 things.

1. Is this embedded preg_replace the best solution? Couldn't think of a better one.

2. $somecode will really be an html-like code with brackets and an undetermined length.
   I had intended to use something like [<*.?>] but was unable to get that to work--any ideas why? (for this test, I used CCC to avoid any display problems.)

Thanks for all your help!

Dan

[Anon]

\w won't match the <, so as soon as it runs up against one the regexp engine will stop trying to match \w*.

If somecode doesn't have any bits in it you want, you should just be able to drop it into the search string like "\w$somecode\w" - adding ()s to get the bits you want for the replacement string. If $somecode might contain things that have special meanings in regexps - ., *, ? and the like - then preg_quote() it first.

Your regexp at the end, I take it you mean "<\w*>"? The backslashes aren't necessary, as < > don't mean anything special to the regexp engine.