How can split on a comma except where it appears in brackets?

Question

I want to split like this:

Before:

TEST_A, TEST_B, TEST_C (with A, B, C), TEST_D

After:

TEST_A
TEST_B
TEST_C (with A, B, C)
TEST_D

How can I split it?

Is the output to be text, HTML, an array or something else? The answer will help you to decide whether to use newline characters ("\n"), HTML breaks, or some other method. — George Cummins, Feb 17 '15 at 13:22
@Rizier123 Yes, I tried. But I can't split except comma in brackets — WebEngine, Feb 17 '15 at 15:37
@Dave I know about regex. but, I can't make a code perfectly works. — WebEngine, Feb 17 '15 at 15:38

score 4 · Accepted Answer · edited May 23 '17 at 11:50

Regex isn’t going to help this time, so you will have to iterate through the characters.

Fact is, regular expressions aren’t very context-aware. For that reason, you can’t use regular expression to parse HTML. This is why we’re better off iterating through the string ourselves.

function magic_split($str) {
    $sets = array('');  // Sets of strings
    $set_index = 0;     // Remember what index we’re writing to
    $brackets_depth = 0; // Keep track if we’re in brackets (or not)

    // Iterate through entire string
    for($i = 0; $i < strlen($str); $i++) {
        // Skip commas if we’re not in brackets
        if($brackets_depth < 1 && $str[$i] === ',') continue;

        // Add character to current list
        $sets[$set_index] .= $str[$i];

        // Store brackets depth
        if($str[$i] === '(') $brackets_depth++;
        if($str[$i] === ')') $brackets_depth--;

        if(
            $i < strlen($str) - 1 && // Is a next char available?
            $str[$i+1] === ',' &&   // Is it a comma?
            $brackets_depth === 0   // Are we not in brackets?
        ) $sets[++$set_index] = '';  // Add new set
    }

    return $sets;
}

$input = 'TEST_A, TEST_B, TEST_C (with A, B, C), TEST_D';
$split = magic_split($input);

So, I add some code[ if($brackets_depth === 0 && $str[$i] === ',') continue; ] And then, It works great. Thank you! — WebEngine, Feb 17 '15 at 15:36
@WebEngine My apologies, you’re completely right. I’ve modified the code if anyone happens to come across this question :) — Tim S., Feb 17 '15 at 17:21

Adam · Answer 2 · 2015-02-17T15:50:20.933

2

You want to match:

a word not containing opening parentheses, nor coma : [^(,]+
an expression between parenthesis: $[^(]+$
- or not... and without returning the match, so it becomes: (?:$[^(]+$)?)
a coma, followed by some space : ,[\s]*

PHP Code:

$ar=preg_split("#([^(,]+(?:\([^(]+\))?),[\s]*#", "$input,", -1,
            PREG_SPLIT_DELIM_CAPTURE |PREG_SPLIT_NO_EMPTY)

Edit: it does not work if you don't have coma outside the parenthesis. you'll have to add an extra coma after $input like modified above.

edited Feb 17 '15 at 15:50

answered Feb 17 '15 at 14:01

Adam

17,838
32
54

It is very great! However, if $input is 'TEST(A,B,C)', It returns 'TEST(' / 'A' / 'B' / 'C)' :( – WebEngine Feb 17 '15 at 15:30
oh you're right. just add an extra coma on the right. – Adam Feb 17 '15 at 15:50

score 1 · Answer 3 · answered Feb 17 '15 at 13:30

The correct solution to this problem will depend on exactly what your specification is for identifying individual elements.

If you expect each one to begin with TEST_, then you could solve it fairly simply with a regular expression:

$input = 'TEST_A, TEST_B, TEST_C (with A, B, C), TEST_D';
$matches = preg_split('/,\s*(?=TEST_)/', $input);

var_dump($matches);

Output:

array(4) {
  [0]=>
  string(6) "TEST_A"
  [1]=>
  string(6) "TEST_B"
  [2]=>
  string(21) "TEST_C (with A, B, C)"
  [3]=>
  string(6) "TEST_D"
}

This splits the string on commas followed by whitespace, using a lookahead assertion test for the presence of TEST_ at the beginning of the next item.

score 1 · Answer 4 · answered Dec 20 '18 at 23:36

You merely need to explode on comma-space and disregard any comma-spaces that are inside of parentheses. (*SKIP)(*FAIL) will consume all parenthetical expressions and dispose of them so that they are not used as delimiters.

Code: (Demo)

$text = 'TEST_A, TEST_B, TEST_C (with A, B, C), TEST_D';

var_export(preg_split('~\([^)]*\)(*SKIP)(*FAIL)|, ~', $text));

Output:

array (
  0 => 'TEST_A',
  1 => 'TEST_B',
  2 => 'TEST_C (with A, B, C)',
  3 => 'TEST_D',
)

How can split on a comma except where it appears in brackets?

Before:

After:

4 Answers4