I just need help... :(

[AV1611]

First, hello to those who may remember me from YEARS ago

 

Anyways, after all these years, I still suck at regex, so here's a real world problem I need help with:

 

Query returns something like this:

 

C1,C2-6,C7,C14,C15-43

 

or:

 

C1,C2-C6,C7

 

or:

 

CX1,CX2-15

 

or:

 

R1,R2-5,R7

 

etc...

 

Anyways, I need to convert to this format: (I will use the last example)

R1,R2,R3,R4,R5,R7

 

I can use array or whatever, I just don't know how to do the regex to do it...

 

And before I get asked for code, here is what I have so far LOL...

 

<?php

$str="R1,R2-5,R7";

 

?>

[JAY6390]

You aren't going to be solving this using regex, at least not all of it. You're going to need to explode the values using the , then us a regex to match the 2-5, then use a for loop to iterate over the values and add them correctly to the array

[.josh]

$string = "C1,C2-6,C7,C14,C15-43";
$string = explode(",",$string);
$list = array();
foreach ($string as $val) {
  preg_match('~([a-z]+)(\d+)(?:-(\d+))?~i',$val,$parts);
  if ( isset($parts[3]) ) {
    for ($c=$parts[2]; $c<=$parts[3]; $c++) {
      $list[] = $parts[1] . $c;
    }
  } else {
    $list[] = $val;
  }
}
$string = implode(',',$list);

 

edit: modified regex to match 1 or more letter prefix

[AV1611]

You aren't going to be solving this using regex, at least not all of it. You're going to need to explode the values using the , then us a regex to match the 2-5, then use a for loop to iterate over the values and add them correctly to the array

I figured I would have to explode the "," then the -... but how do I get the "alpha" character, seeing it may be 1-3 characters long? i.e. R, C, CX, etc...?

Maybe something that does "grab the beginning of the string until you encounter a number"? How do I do that?

[JAY6390]

[^\d]+ will match everything before the numbers

[AV1611]

Almost works...

 

$str = CX1,CX2-6,CX7

 

yields: CX1,X2,X3,X4,X5,X6,CX7

 

Seems to work other than that...

 

[.josh]

if you are referring to the code I posted, look at the edit. I noticed afterwards that the letter prefix could be more than 1 char so i adjusted the regex.

[AV1611]

if you are referring to the code I posted, look at the edit. I noticed afterwards that the letter prefix could be more than 1 char so i adjusted the regex.

 

My hero

[AV1611]

Oops... one last problem...

 

The data is coming from a VERY old database, and the data syntax is not always the same. The following is possible:

$str="CX1,CX2-6,CX7,CX9-CX11";

 

They did CX2-6, then later CX9-CX11. That seems to be a small issue as well... I am so sorry...

[.josh]

geez man, you're killin' me. Are there any other surprises or "gotchas" you need to mention?

[JAY6390]

<?php

$str = "R1,R2-5,R7";
$parts = explode(',', $str);
$out = array();
$regex = '~^([^\d]+)(\d+)-([^\d]*)(\d+)$~';

foreach($parts as $part) {
if(preg_match($regex, $part, $matches)) {
	$range = range($matches[2], $matches[4]);
	foreach($range as $number) {
		$out[] = $matches[1] . $number;
	}
} else {
	$out[] = $part;
}
}

echo '<pre>' . print_r($out, true) . '</pre>';

?>

 

That's how I would do it. Should work for your last example too

[.josh]

good job Jay, but IMO you unnecessarily complicated it.  No need to group that additional ([^\d]*) and use extra matched index.  Also that would technically match non-alphanumeric chars.  Probably not a concern but may as well be more efficient and explicit since it is assumed that the prefix are alpha only.

 

Also, because I was exceedingly bored, I one-lined it:

 

$string = "CX1,CX2-6,CX7,CX9-CX11";
$string = preg_replace_callback('~([a-z]+)(\d+)(?:-[a-z]*(\d+))?~i',create_function('$m','return (isset($m[3]))?($m[1].implode(\',\'.$m[1],range($m[2],$m[3]))):$m[0];'),$string);
echo $string;

 

output:

 

CX1,CX2,CX3,CX4,CX5,CX6,CX7,CX9,CX10,CX11

[JAY6390]

ha yeah I like to keep my groups just so I know where everything is tbh, you are right though it was unnecessary. Love the one liner!

[AV1611]

You guys are awesome... I was trying to get the final answer (please don't laugh too loud... it would have worked... but I didn't get part of it right...) anyways, here's my crappy approach. Remember, I don't know regex, but I can improvise LOLOL

 

$string = "CX1,CX2-6,CX7,CX9-CX11";

preg_match('~([a-z]+)(\d+)(?:-(\d+))?~i',$string,$parts);

$prefix = $parts[1];

$string=str_replace(" ","",$string);

$string=str_replace($prefix,"",$string);

$string = explode(",",$string);

$c=count($string);

$i=0;

while($i < $c){

$s=explode("-",$string[$i]);

$ct=count($s);

$ci=0;

while($ci < $ct){

echo $prefix . $s[$ci] . "<br />\r\n";

$ci++;

}

$i++;

}

 

[Philip]

you unnecessarily complicated it

$string = preg_replace_callback('~([a-z]+)(\d+)(?:-[a-z]*(\d+))?~i',create_function('$m','return (isset($m[3]))?($m[1].implode(\',\'.$m[1],range($m[2],$m[3]))):$m[0];'),$string);

[.josh]

Okay you got me phil, that one-liner is way more complicated from a readability perspective.  But to be fair, I was speaking from a "more steps than necessary" perspective.

 

But anyways..

 

don't be jealous!