[SOLVED] preg_match a character, except where it's wrapped in braces

[Mark Baker]

I'm not particularly good with regular expressions, but I need to do a preg_match to see if a string contains an instance of a particular character where it isn't wrapped in square braces.

 

e.g. If I was testing for the character 'd'

"dd-MMM-yyyy"

    should return a TRUE because d is in the string, and there are no square braces

"[$-F800]dddd\,\ mmmm\ dd\,\ yyyy"

    should return a TRUE because there is a d in the string outside of the square braces

"[Red]dddd\,\ mmmm\ dd\,\ yyyy"

    should return a TRUE because there is a d in the string outside of the square braces, even though there is also a d within the square braces

"[Red]#,##0.00;[Green]-#,##0.00"

    should return a FALSE because there is no d in the string outside of the square braces

"#,##0.00"

    should return a FALSE because there is no d in the string irrespective of the fact that there are no square braces either

 

Previously, I've simply been using stripos to check whether a 'd' exists, but returning a false if there are square braces as well, but this is no longer an adequate check, so preg_match seemed a better more flexible alternative

 

 

I also have a loop which executes this script for several characters in an array, 'd' being just one such character. If any of these characters exist (outside of square braces), then a True is returned. False is only return if none of the characters in the array exist (outside of the square braces) in the tested string.

I'm sure a preg_match could also allow me to test for any of the characters in the array in a single call, rather than needing to loop.

 

However, I'm way out of my depth with regexp. Any help would be much appreciated

[Mark Baker]

Looks like some of my examples are setting message formatting

"dd-MMM-yyyy"
    should return a TRUE because d is in the string, and there are no square braces
"[$-F800]dddd\,\ mmmm\ dd\,\ yyyy" 
    should return a TRUE because there is a d in the string outside of the square braces
"[Red]dddd\,\ mmmm\ dd\,\ yyyy"
    should return a TRUE because there is a d in the string outside of the square braces, even though there is also a d within the square braces
"[Red]#,##0.00;[Green]-#,##0.00"
    should return a FALSE because there is no d in the string outside of the square braces
"#,##0.00"
    should return a FALSE because there is no d in the string irrespective of the fact that there are no square braces either

 

Currently I'm using a preg_replace first to strip out the [] and their content before doing the preg_match

$formatValue = preg_replace('/\[(.*?)\]/','',$formatValue);
foreach ($possibleCharacters as $possibleCharacter) {
if (preg_match('/'.$possibleCharacter.'/i',$formatValue)) {
	return True;
}
}
return False;

[nrg_alpha]

For this example, I just lumped some of your samples into an array and check from there. So, if I understand this correctly

 

$str = array('"dd-MMM-yyyya"', '"[$-F800a]dddd\,\ mmmm\ dd\,\ yyyy"', '"#,##0.00;-#,##0.00"', '"[dda]-MMM-yyyy"', '"dd\yyyy\a[d]\mmmm"');
$match = NULL;
foreach($str as $val){
   if(!preg_match('#\[.*?d.*?\]#', $val) && strpbrk($val, 'd')){
      $match[] = $val;
   }
}
echo '<pre>'.print_r($match, true);

 

Output:

Array
(
    [0] => "dd-MMM-yyyya"
    [1] => "[$-F800a]dddd\,\ mmmm\ dd\,\ yyyy"
)

 

Am I on the right track?

[Mark Baker]

I see what you're trying to do: if there isn't a 'd' in [], but there is still a 'd' somewhere in the string, then accept.

But I need a positive if there is a 'd' in the non-bracketed part of the string, even if there is a bracketed 'd' as well; so I need a valid match for "dd\yyyy\a[d]\mmmm" as well.

 

My preg_replace to strip the brackets and their content followed by preg_match to test for the character does work as expected... I was hoping it would be possible to do the entire test in a single preg_match though.

[sasa]

try

<?php
$test[] = "dd-MMM-yyyy";
//    should return a TRUE because d is in the string, and there are no square braces
$test[] = "[$-F800]dddd\,\ mmmm\ dd\,\ yyyy"; 
//    should return a TRUE because there is a d in the string outside of the square braces
$test[] = "[Red]dddd\,\ mmmm\ dd\,\ yyyy";
//    should return a TRUE because there is a d in the string outside of the square braces, even though there is also a d within the square braces
$test[] = "[Red]#,##0.00;[Green]-#,##0.00";
//    should return a FALSE because there is no d in the string outside of the square braces
$test[] = "#,##0.00";
$char = 'd';
$pattern = '/(^|\])[^\[]*'.$char.'/';
foreach ($test as $t){
echo $t,' -> ', preg_match($pattern,$t) ? 'TRUE' : 'FALSE', "<br />\n";
}
?>

[Mark Baker]

Thanks sasa

That works a treat.

 

I can even test for all possible variations of $char by setting it as '[dmyhis]', which will return a true if any of those characters are in the string being tested, which is what I needed; so I don't need to loop through my array of individual valid characters any more.

[sasa]

it doesn't work if you have nested brace ('[a[bc]d]')

[Mark Baker]

it doesn't work if you have nested brace ('[a[bc]d]')

Appreciate that, but that shouldn't ever occur. I know that my string to search should always be cleanly generated, because I have an implicit trust in the developers of MS Excel, and of those applications capable of writing Excel format files :-)

 

Thanks again for the solution.