'How do I match this pattern using preg_match in PHP?

I'm writing a simple quiz engine in PHP and supply the question text in this format

question|correct/feedback|wrong/feedback|wrong/feedback

There can be as many wrong/feedback options as necessary. I want to use preg_match to return the results so I can display them. For instance:

q|aaa/aaa|bbb/bbb|ccc/ccc

...should return...

array(
 0 => q|aaa/aaa|bbb/bbb|ccc/ccc
 1 => q
 2 => aaa/aaa
 3 => bbb/bbb
 4 => ccc/ccc
)

So, far I've got this regular expression which matches the question and the correct/feedback combination...

([^\|]+)\|([^\/]+\/[^\|$]+)

...but I have no idea how to match the remaining wrong/feedback strings

phpregexpreg-match


Solution 1:[1]

You can also use the "glue" feature in your pattern with preg_match_all, this way it's possible to check if the syntax is correct and to extract each part at the same time.

The glue feature ensures that each match follows immediately the previous match without gap. To do that I use the A global modifier (Anchored to the start of the string or the next position after the previous match).

$s = 'q|aaa/aaa|bbb/bbb|ccc/ccc';

$pat = '~ (?!\A) \| \K [^|/]+ / [^|/]+ (?: \z (*:END) )? | \A [^|/]+ ~Ax';

if ( preg_match_all($pat, $s, $m) && isset($m['MARK']) ) {
    $result = $m[0];
    print_r($result);
}

I use also a marker (*:END) to be sure that the end of the string is well reached despite of the pattern constraints. If this marker exists in the matches array, it's a proof that the syntax is correct. Advantage: you have to parse the string only once (you don't even need to check the whole string syntax in a lookahead assertion anchored at the start of the string).

demo

If you want the whole question as first item in the result array, just write:

$result = array_merge([$s], $m[0]);

Solution 2:[2]

So, after the advice, I've decided to use preg_match to check the syntax and then explode to split the string.

This regex seems to match the string format up until any mismatch occurs.

^[^\|/]+(?:\|[^\|/]+/[^\|/]+)+

If I check that the length of the match is the same as the original string I think this will tell me the syntax is correct. Does this sound feasible?

Sources

This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.

Source: Stack Overflow

Solution Source
Solution 1
Solution 2 MrMills

Related Questions