REGEX Help ...

jtown

I need some help with php and regular expressions.

Say I have a string "http://www.phpbuilder.com/board/newthread...". How do I do work on this string and only return "http://www.phpbuilder.com"?

I know how to do this in Perl, but doesn't seem like a 1 step thing in PHP.

Thanks and Merry Xmas.

sfullman


if(preg_match('/^(http:\/\/)*([-a-z0-9]+(\.[-a-z0-9]){1,})/i',$url,$match)){
   $domain=$match[2]; //2nd parenthesis
}
/***
note that the domain will have a-z, 0-9 and a dash possibly.  The {1,} allows for complex domains like ci.san-marcos.tx.us which might have multiple periods in them, however there will always be at least one period (except for localhost :-)
HTH
***/

laserlight

Actually, have you considered using [man]parse_url/man?

jtown

laserlight wrote:
Actually, have you considered using [man]parse_url/man?

Thanks! Seems to work for my purposes.

I guess I kinda figured out how to do the perl regexp in php too ...

In Perl:

$url =~ /^{(http:\/\/.*?)\//);}

$parsed_url = $1;

In PHP this might work (didn't try):

if(preg_match('/^{(http:\/\/.*?)\//',$url,$match)){}
$parsed_url=$match[1];
}

RyanJones

A bit offtopic I know but if you are interested in using RegExps then here may be a good plact too start :-)

http://www.regexplib.com

Cheers,

Ryan Jones

Phantazm

Someone take a stab at this and potentially help me clean it up. It is used to parse out the WPFileName and WPFileType (along with other things, as you'll see) from WP51 files. As sloppy as it is, it works.. MOST of the time... I'm having problems with a portion of it, however.. I'll explain.

    // Constants
    $wpdatetime = '[0-9]{2}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}[ap]';
    // File
    $filedata = file_get_contents($file);
    //WP Data Extractor
    $pattern = '/('.$wpdatetime.')\x20{10}\x00(.{68})(.{20})\x00{7}\xFF(.*\xA9\x20([0-9]{0,5}))?/i';
    if(!preg_match($pattern, $filedata, $matchdata[]))
    {
      echo "$file - Missed it!<br>";
    }

Explaining it (and my likely mixed up logic at 3:47 am!).. WordPerfect 5.1 saves files with the following predictible patterns:

10 spaces (\x20{10}) -->
1 NULL (\x00) -->
88 Characters I want in two variables (this works!) -->
7 NULLS (\x00{7}) -->
1 hFF (\xFF) -->

-- here is where the trouble starts --

Any number of unknown characters... 90-120 or so ... followed by some data I want with the pattern...

\xA9\x20([0-9]{0,5})

How in the h*** do I "skip" over the junk in some way I am not thinking of... I have ONE file (yes one.. buth there is something wrong in the regex) that my regex does not work on, despite it seemingly (here's MY error) matching the pattern.

Ideas??? Clean-ups?? 😛 😕

(.*\xA9\x20([0-9]{0,5}))?

That extra set of ()? enclosing this is JUST to get the WHOLE to work if it somehow MISSES the particular data on the given file I'm having problems with!

RyanJones

Too skip a portion, use a pattern that matched anything, try b[/b] :-)

Cheers,

Ryan Jones

Phantazm

Okay... definantly better.. but I am, for some reason, still not getting a match in ONE stinking file. (No change from results of my regex).

New regex is:

    $wpdatetime = '[0-9]{2}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}[ap]';
    $pattern = '/('.$wpdatetime.')\x20{10}\x00(.{68})(.{20})\x00{7}\xFF(.+?)\xA9\x20([0-9]{0,5})/i';

Is there anything in a file that would "crash" a regex as to NOT find the previously mentioned pattern... namely everything AFTER the (.+?) --- ?

If I run the

\xA9\x20([0-9]{0,5})

through ANOTHER preg_match, I get the data I want. Why do it in TWO steps when it should be done in one.. right?

Would some actual data (posted in hexadecimal, I suppose) from the file be helpful. I, unfortuantly, cannot post the offending file in its entirety, as it is a Medical Record (I'm parsing medical transcription records and snagging the PAtient Name, Patient Number, etcetera...)

Thanks again.

Phantazm

One More....

    $pattern = '/('.$wpdatetime.')\x20{10}\x00(.{68})(.{20})\x00{7}\xFF/i';
    $pattern2 = '/\xA9\x20([0-9]{0,5})/i';

$pattern - this finds the first portion of data I want just fine
$pattern2 - this ALWAYS finds the second small chunk of data

Whatever I try to put them TOGETHER fails. There is one file, that quite obviously follows the pattern of BOTH (as it works independently), yet I CANNOT get to work then I "joint" the regexs with something like (.*) or better yet (.+?)

The second pattern ALWAYS comes after the first.

Ideas?

Phantazm

solved

$pattern = '/('.$wpdatetime.')\x20{10}\x00(.{68})(.{20})\x00{7}\xFF(.+?)\xA9\x20([0-9]{0,5})/s';

One *&$% character! Isn't that ALWAYS all it is???