regexp problem - please help?

[iceomnia]

Hi Everyone,

I need the following with a regular expression and I also need to know which to use from regexp_match and eregi_match or whatever.

The following:

Soundlab
Sound lab
SoundLab
Sound Lab
Sound-lab
sound-Lab
soundlab

all match the word 'Soundlab'

And for another example:

Audio-technica
Audiotechnica
audio Technica
Audio Technica
Audio Technica

all match the word/phrase 'Audio Technica'

how would I use regexp to match these phrases, it has to be universal with the amount of words from 1 word to 10 words so I can loop through manufacturers and find duplicates.

Can anyone help?

[iceomnia]

Hello?

[bradgrafelman]

The immediate problem is that given input "soundlab", how is PHP supposed to know where one word ends and the second/third/etc. begins? Given "sound lab", it'd be quite easy to do what you're suggesting.

I can't think of an easy solution, except for perhaps using a full text search feature built into DB's such as MySQL?

[iceomnia]

Yes, that's what I mean:

if(eregi('sound lab') == 'soundlab')
{
  echo 'this is a match';
}

[iceomnia]

Obviously. the aboce code is wrong, but thats my point...?

[bradgrafelman]

if(preg_match('/sound[ -]?lab/i', 'Sound-Lab'))
    echo 'This is a match.';

[iceomnia]

Fantastic, Thanks... That is the exact answer I was looking for, however, I need it be generic for any of many words in the first post.

[Lillefix]

<?
$string = "Soundlab";
$compare = "Sound-Lab";
$remove = array(' ','-');

$trimmed = str_replace($remove, "", $compare);

if (strtolower($compare) == strtolower($string))
{
	echo "Duplicate";
}
?>

[iceomnia]

Thanks...

But I have encountered a new problem now...

The compare string count have any one of the following chars:

& - With a space at either side
-
+
&

This is why I wanted to use a regexp, any ideas?

Regexp is one thing I cannot get my head around!

[laserlight]

Instead of defining what to remove, define what to keep, then remove everything else. You can use preg_replace() instead of str_replace() for this. Note that strcasecmp() can replace the use of strtolower().

[iceomnia]

Hi everyone again, looks like I'm back to this...

The manufacturers name may be any one of the following examples:

Soundlab
Allen & Heath
Pioneer
Blackberry
Nokia

The below ARE different manufacturers:

Sony Erricson
Sony

And I need the following to work for all those manufacturer names:

When there's a space: Sound Lab (would still match 'Soundlab'),

When there's a hyphen: Sony-Erricson ('Sony Erricson')

When there's a dash instead of a & symbol ('Allen-Heath')

a space insead of a & symbol: ('Allen Heath') would still match 'Allen & Heath',

Also: where a single word would NOT match something with more than one word for example:

'Sony Erricson' would NOT match 'Sony'; but 'SoNy' would match 'Sony'.

I know this is getting really complicated, but my regexp skills are non-existent.

I really need ONE single pattern to do all this for me, would anyone be willing to write a regexp pattern for preg_match or ereg_match or whatever that would help me with this?

[bradgrafelman]

I would say remove all non-alphanumeric characters (e.g. spaces, dashes, ampersands, etc.) from both the search keyword(s) as well as the manufacturer names and then do a case-insensitive match.

[iceomnia]

Do you have an example?