Explode string when not between ()

Question

I'm trying to explode an string by comma's: , but only when it's not between parenthesis (...)

Here's some code to show what I mean:

$string = "K.VisueelNr,IFNULL(P.Partijnaam, 'Leeg gemeld') AS Partijnaam,R.Ras,M.Maat,DATE_FORMAT(KS.Timestamp, '%d-%c-%Y') AS Datum,H.Handelshuis,KS.Actie";

print_r(str_getcsv($string, '()'));

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

Which will output:

Array
(
    [0] => K.VisueelNr,IFNULL
    [1] => P.Partijnaam, 'Leeg gemeld') AS Partijnaam,R.Ras,M.Maat,DATE_FORMAT
    [2] => KS.Timestamp, '%d-%c-%Y') AS Datum,H.Handelshuis,KS.Actie
)
Array
(
    [0] => K.VisueelNr
    [1] => IFNULL(P.Partijnaam
    [2] =>  'Leeg gemeld') AS Partijnaam
    [3] => R.Ras
    [4] => M.Maat
    [5] => DATE_FORMAT(KS.Timestamp
    [6] =>  '%d-%c-%Y') AS Datum
    [7] => H.Handelshuis
    [8] => KS.Actie
)

But I'm looking for an output like this:

Array
(
    [0] => K.VisueelNr
    [1] => IFNULL(P.Partijnaam, 'Leeg gemeld') AS Partijnaam
    [2] => R.Ras
    [3] => M.Maat
    [4] => DATE_FORMAT(KS.Timestamp, '%d-%c-%Y') AS Datum
    [5] => H.Handelshuis
    [6] => KS.Actie
)

Here's an PHP Sandbox fiddle to play around

Maybe it has to be done with preg_split, but I don't know how the regex would look like then...

Sumerlin

Answer 1

You need to use preg_split which splits the input according to the given regex.

$words = preg_split('~,(?![^()]*\))~', $str);
print_r($words);

DEMO

Explanation:

• , matches all the commas, only if it's not
• followed by any char but not of ( or ), zero or more times
• and further followed by a closing bracket.

If you change (?! to (?=, it does the opposite of matching all the commas which are present inside the brackets.

Hallvard

Thank you very much. Its works indeed as expected. And thank you very much for the demo link also. The information provided by the explanation is very helpfull – Sumerlin

@Sumerlin Note that this method is very fragile. It breaks when there are nested function calls. – Wakefield

@Wakefield luckily that's not the case. But now I'm curious. Can you point me to somewhere with an explanation for this problem? – Sumerlin

@Mathlight: Here is an example func(arg1, func(arg2)) as field. You can't expect it to work correctly in general when you only look at some characters in front of a ,. – Wakefield

Answer 2

Needed a solution that works with nested brackets and came up with this:

$delimiter = ',';
$string = "K.VisueelNr,IFNULL(P.Partijnaam, 'Leeg gemeld') AS Partijnaam,func(arg1, func(arg2)) as field";

$array = explode($delimiter, $string);
$new_array = [];
$helper_string = '';
$count1 = 0;
$count2 = 0;
foreach ($array as $str) {
    $count1 += substr_count($str, '(');
    $count2 += substr_count($str, ')');
    if (strlen($helper_string) > 0) {
        $helper_string .= $delimiter;
    }
    $helper_string .= $str;
    if ($count1 === $count2) {
        $new_array[] = $helper_string;
        $helper_string = '';
        $count1 = 0;
        $count2 = 0;
    }
}
// Fallback in case of broken brackets
if (strlen($helper_string)) {
    $new_array[] = $helper_string;
}
print_r($new_array);

Result:

Array
(
    [0] => K.VisueelNr
    [1] => IFNULL(P.Partijnaam, 'Leeg gemeld') AS Partijnaam
    [2] => func(arg1, func(arg2)) as field
)

Butlery