REGEX Help (I think?)

[macwise]

Ok, my first post here (earlier today) was a huge help, so here's a second...

 

I have the following PHP code:

 

$name = $row['NAME'];
$nameTrimmed = $name;

if ((strlen($nameTrimmed) > 20) && (strlen($nameTrimmed) > 1)) { 
  $whitespaceposition2 = strrpos($nameTrimmed," ",5); 
  $nameTrimmed = substr($nameTrimmed, 0, $whitespaceposition2); 
  echo $nameTrimmed . "...";
}

if ((strlen($name) <= 20) && (strlen($name) > 0)) {
  echo $name;
}	

 

This is working as expected.  However, many of the "name" fields in the DB contain ™ and ® strings, which inflate the count of each product title.  Since I'm using the strrpos to limit the length of the longer product names (due to limited space in the layout), I would like to show as many full titles as possible.  Many are being cut prematurely short due to the extra characters from the symbols. 

 

Do I need to run all this through a regex filter before I count it?  If so, can anyone share an example of code I could try? Regex has never been my weak point, let alone my strong point.    Or is there something else I haven't thought of? 

 

Any help is greatly appreciated!

[RussellReal]

try this regex..

 

preg_replace("/&[a-z]{2,5};/i",'X',$title);

 

that SHOULD work..

 

however.. y ou could just pass over with html_entity_decode

[btherl]

If you really have "®" in your db, you might need to html_entity_decode() twice to get the actual character.  Keep in mind that ® and ™ are not going to map to ascii characters though ..

 

Instead of decoding, you might want to just replace "&...;" sequences with a single dummy charater, so you can get the length count correct.  Then use the original string for displaying.

[RussellReal]

If you really have "®" in your db, you might need to html_entity_decode() twice to get the actual character.  Keep in mind that ® and ™ are not going to map to ascii characters though ..

 

Instead of decoding, you might want to just replace "&...;" sequences with a single dummy charater, so you can get the length count correct.  Then use the original string for displaying.

 

thats what I do in my first example with "X" but yes he does make a point about the ascii characters, I'm not a real guru with charsets or anything I just know regex n stuff

[macwise]

Awesome! Thanks!

 

Here's what I ended up with (is this totally hacked looking or did I do this in an acceptable way?):

$name = $row['NAME'];
$nameTrimmed = $name;

if ((strlen(preg_replace("/&[a-z]{2,5};/i",'X',$name)) > 16) && (strlen($name) > 1)) { 
  $whitespaceposition2 = strrpos($name," ",5); 
  $nameTrimmed = substr($name, 0, $whitespaceposition2); 
  echo $nameTrimmed . "...";
} else {
  echo $name;
}	

 

Seems to work great.  Thanks again—I suck at regex!

[btherl]

That looks right to me.  As long as you only need a single level of decoding.  You might want to check the result of that preg_replace() to make sure it is what you expect it to be.

 

Russel, if you're replacing with X then you don't need to worry about character sets luckily   Unless there are some html entities that appear as more than one character, that's a possibility .. but I can't think of any.