[RESOLVED] preg_replace insert hyphen -- ABC123 -> ABC-123, A1 -> A-1, etc.

[dalecosp]

I'm up against a brick wall on this one ... coming into work with the top down seems to have addled my already troubled mind

My "in string" is of this format:

$std_regex1="%[\w]{1,3}\d{1,5}%";   //A1, ABC12345 

I want to separate the alpha character(s) from the numerics with a hyphen - the desired "out string" is "A-1" in the first example case, and "ABC-12345" in the second.

So far, I've drawn a blank on this with preg_replace() and was actually considering looping through the entire string to prepend a hyphen to the first instance of a numeric character, but I'm already way too expensive to scale this.

A good 1/2 hour of searching here and the big G isn't helping, either. Maybe more breakfast/coffee?

Any pointers are appreciated ... I'm sure I'm just being thick as usual.

[Derokorian]

Give this a try:

$array = array('A1','ABC12345');
foreach( $array as $k => $v ) {
   $array[k] = preg_replace('/(\w{1,3})(\d{1,5})/','\1-\2',$v);
}
var_dump($array); 

[dalecosp]

Thanks! It's close, but different test cases fail. Did I not give adequate test cases? The regex will also match things like ab2345, yz33, etc.

[574] Wed 24.Oct.2012 11:28:43 [kevin@rootbsd][~] cat foo; php foo; cat bar; php bar
<?php
$array = array('AB2345','yz33');
foreach( $array as $k => $v ) {
   $array[$k] = preg_replace('/(\w{1,3})(\d{1,5})/','\1-\2',$v);
}

var_dump($array);
array(2) {
  [0]=>
  string(7) "AB2-345"
  [1]=>
  string(5) "yz3-3"
}

<?php
$array = array('a1','abc123');
foreach( $array as $k => $v ) {
   $array[$k] = preg_replace('/(\w{1,3})(\d{1,5})/','\1-\2',$v);
}
var_dump($array);
array(2) {
  [0]=>
  string(3) "a-1"
  [1]=>
  string(7) "abc-123"
}

If it could get the hyphen one character sooner in the other cases it looks golden ... though I need to throw a few thousand test cases at it to be certain.

[Derokorian]

Well, then do this instead:

<?php

$array = array('A1','ABC12345','AB2345','yz33');
foreach( $array as $k => $v ) {
   $array[$k] = preg_replace('/([a-z]{1,3})(\d{1,5})/i','\1-\2',$v);
}
var_dump($array);

Output:

>php test2.php
array(4) {
  [0]=>
  string(3) "A-1"
  [1]=>
  string(9) "ABC-12345"
  [2]=>
  string(7) "AB-2345"
  [3]=>
  string(5) "yz-33"
}

[dalecosp]

Blargh, yes, indeed. Just did it for myself. Thanks. I need a lunch break!

[dalecosp]

Shouldn't have been using "\w" in the first place on this one, I think

It was imported from another function where it needed to be less restrictive.

[johanafm]

I'd recommend something I find a lot simpler in this case. Create a pattern consisting of a looh behind assertion for a-zA-Z and a look ahead assertion for 0-9. That is, you'll be matching the spot where a letter is followed by a didigt, and replace that spot with a -. That there isn't an actual character in that particular spot doesn't matter, it works anyway.

$arr = array(
	'abc123',
	'ZY99',
	'ASDDQWE4564',
	'ASD',
	'123',
);

$re = '#(?<=[a-zA-Z])(?=[0-9])#';

foreach ($arr as $s)
{
	$out = preg_replace($re, '-', $s);
	echo $out.'<br>';
}

[Derokorian]

Stop being fancy johanafm. :P