Why is preg_match_all returning two matches?

Question

I am trying to identify if a string has any words between double quotes using preg_match_all, however it's duplicating results and the first result has two sets of double quotes either side, where as the string being searched only has the one set.

Here is my code:

$str = 'Test start. "Test match this". Test end.';
$groups = array();
preg_match_all('/"([^"]+)"/', $str, $groups);
var_dump($groups);

And the var dump produces:

array(2) {
    [0]=>
    array(1) {
        [0]=>
        string(17) ""Test match this""
    }
    [1]=>
    array(1) {
        [0]=>
        string(15) "Test match this"
    }
}

As you can see the first array is wrong, why is preg_match_all returning this?

Spa

Leave out the brackets, or add some brackets, read the manual and you'll work it out – Fossilize

Please define wrong. How is that not matching up your expectations that php.net/preg_match_all gives? – Conservative

As you don't define wrong, it could be also that you're confusing the output of var_dump here? – Conservative

Yes, I'm confusing var_dump. I didn't realise that preg_match_all recorded with and without the capture, and that it was var_dump adding quotes around the strings. The two lots of double quotes is what I thought was wrong. – Spa

Answer 1

Hi if your are using print_r instead of vardump you will see the differences in a better way.

Array
(
    [0] => Array
        (
            [0] => "Test match this"
        )

    [1] => Array
        (
            [0] => Test match this
        )

)

The first contains whole string and the second is your match.

Photocurrent

Right, didn't realise that preg_match_all returned both, thank you. – Spa

Answer 2

It returns 2 elements because:

Element 0 captures the whole matched string

Elements 1..N capture dedicated matches.

PS: another way of expressing the same could be

(?<=")[^"]+(?=")

which would capture exactly the same but in that case you don't need additional capturing group.

Demo: http://regex101.com/r/lF3kP7/1

Stud

This doesn't produce the results I expect, it includes the end string - eval.in/210849 – Spa

@Styphon: hint: (?=") – Stud

@Avinash Raj: var_dump encloses all string literals in double quotes. It's not what a question is about though. – Stud

@Avinash Raj: that's right, and print_r does not solve the "issue". It's about understanding how preg_match_all works, not about print_r vs var_dump – Stud

@Avinash Raj: that's not what OP asked in the question though. I'm sorry, but I cannot read OP's minds. – Stud

@Avinash Raj: for me it looks like OP was confused about 2 items instead of one. I could not realize someone would be confused that a string literal is delimited in quotes, it's just... unbelievable :-S – Stud

I was asking why there was two results, one with quotes and one without, but that was me being confused by var_dump and then not understanding how preg_match_all worked. – Spa

@Styphon: sorry then. I'm used to thinking that an OP with 100+ answers checks the documentation first. I was so wrong :-D – Stud

I had checked the documentation, I just misunderstood it. It's easy enough to do if you're not used to using something. – Spa

@Styphon: Fair enough (I would try to use the function I do not understand on multiple inputs though before I ask a question. var_dump('foo') would demonstrate its behaviour) – Stud

Point taken. I'll make sure to do that in future. – Spa

Answer 3

Remove the parenthesis. you can write the pattern as '/"[^"]+"/'

Demarcate

Answer 4

This is because you're using group matching. take the parentheses out of your pattern and you'll get one array back. Something like:

preg_match_all('/\"[^"]+\"/', $str, $groups);

Dessert

Why are you escaping the quotes? – Toluidine