how to remove only real duplicates from preg_match output?

问题

Yes I know array_unique function, but the thing is that match might have a legitimate duplicates in my search term for example:

$str = "fruit1: banana, fruit2: orange, fruit3: banana, fruit4: apple, fruit5: banana";
preg_match("@fruit1: (?<fruit1>\w+), fruit2: orange, fruit3: (banana), fruit4: (?<fruit4>apple), fruit5: (banana)@",$str,$match);
array_shift($match); // I dont need whole match
print_r($match);

output is:

Array
(
    [fruit1] => banana
    [0] => banana
    [1] => banana
    [fruit4] => apple
    [2] => apple
    [3] => banana
)

So the only keys that are real duplicates are [0] and [2] but array_unique gives:

Array
(
    [fruit1] => banana
    [fruit4] => apple
)

回答1:

This is my solution for your problem:

unset($matches[0]);
$matches = array_unique($matches);

回答2:

I found it myself, solution is a while loop that deletes subsequent key is one that it is at is not numerical:

while (next($match) !== false) {
  if (!is_int(key($match))) {
    next($match);
    unset($m[key($match)]);
  }
}
reset($match);

来源：https://stackoverflow.com/questions/13333431/how-to-remove-only-real-duplicates-from-preg-match-output