Parse string with regex and get desired output

Question

I want to parse this string

[[delay-4]]Welcome! [[delay-2]]Do you have some questions for us?[[delay-1]] Please fill input field!

I need to get something like this:

[
    [0] => '[[delay-4]]Welcome!',
    [1] => '[[delay-2]]Do you have some questions for us?',
    [2] => '[[delay-1]] Please fill input field!
];

String can also be something like this (without [[delay-4]] on beginning):

Welcome! [[delay-2]]Do you have some questions for us?[[delay-1]] Please fill input field!

Expected output should be something like this:

    [
        [0] => 'Welcome!',
        [1] => '[[delay-2]]Do you have some questions for us?',
        [2] => '[[delay-1]] Please fill input field!
    ];

I tried with this regex (https://regex101.com/r/Eqztl1/1/)

(?:\[\[delay-\d+]])?([\w \\,?!.@#$%^&*()|`\]~\-='\"{}]+)

But I have problem with that regex if someone writes just one [ in text, regex fails and if I include [ to match I got wrong results.

Can anyone help me with this?

What exactly do you want to match? You current patter will give you that output 3v4l.org/MvSiH Do you want to match all consecutive square brackets even when they are unbalanced? Perhaps like this regex101.com/r/g5PlFl/1 — The fourth bird
– The fourth bird, Commented Jul 26, 2019 at 12:03
you regex working fine for both string:- 3v4l.org/PWDK9 and 3v4l.org/MvSiH so what problem you are facing now? — Death-is-the-real-truth
– Death-is-the-real-truth, Commented Jul 26, 2019 at 12:06

Dean Taylor · Accepted Answer · 2019-07-26 13:33:27Z

2

Two simpler actions might be the route to get the result:

$result = preg_replace('/\s*(\[\[delay-\d+]])/i', "\n$1", $subject);
$result = preg_split('/\r?\n/i', $result, -1, PREG_SPLIT_NO_EMPTY);

Can be seen running here: https://ideone.com/Z5tZI3 and here: https://ideone.com/vnSNYI

This assumes that newline characters don't have special meaning and are OK to split on.

UPDATE: As noted in the comments below it's possible with a single split.

$result = preg_split('/(?=\[\[delay-\d+]])/i', $subject, -1, PREG_SPLIT_NO_EMPTY);

But there are possible issues with zero-length matches and regular expressions, you would have to do your own research on that.

edited Jul 26, 2019 at 13:33

answered Jul 26, 2019 at 13:04

Dean Taylor

42.4k3 gold badges33 silver badges53 bronze badges

Sign up to request clarification or add additional context in comments.

6 Comments

Casimir et Hippolyte Over a year ago

What if the string contains line-breaks? It can be done with only one preg_split with your first pattern if you change your capture group with a lookahead. ] isn't a special character.

Dean Taylor Over a year ago

Sure you can - but then you have to be aware of the issues around zero-length matches which is a complex issue to have to explain.

Casimir et Hippolyte Over a year ago

What kind of issues are you speaking about?

Dean Taylor Over a year ago

Zero-length matching issues, some of which are detailed here: regular-expressions.info/zerolength.html Note it's unclear how this is effects preg_split.

Casimir et Hippolyte Over a year ago

3v4l.org/j3IeT : By force of habit I have never see it as an issue! Whatever, since you use the PREG_SPLIT_NO_EMPTY flag, it isn't a problem.

|

The fourth bird · Accepted Answer · 2019-07-26 12:43:21Z

1

In your pattern

(?:[[delay-\d+]])?([\w \,?!.@#$%^&*()|`]~-='\"{}]+)

there is no opening [ in the character class. The problem is that if you add it, you get as you say wrong results.

That is because after matching after matching delay, the character class in the next part which now contains the [ can match the rest of the characters including those of the delay part.

What you could do is to add [ and make the match non greedy in combination with a positive lookahead to assert either the next match for the delay part or the end of the string to also match the last instance.

If you are not using the capturing group and only want the result you can omit it.

(?:\[\[delay-\d+]])?[\w \\,?!.@#$%^&*()|`[\]~\-='\"{}]+?(?=\[\[delay-\d+]]|$)

Regex demo | Php demo

edited Jul 26, 2019 at 12:43

answered Jul 26, 2019 at 12:37

The fourth bird

165k16 gold badges61 silver badges75 bronze badges

Comments

Andreas · Accepted Answer · 2019-07-26 12:06:13Z

1

You can do that without regex too.

Explode on [[ and loop the array. If the start of the item is "delay" then add [[

$str = '[[delay-4]]Welcome! [[delay-2]]Do you have some questions for us?[[delay-1]] Please fill input field!';

$arr = array_filter(explode("[[", $str));

foreach($arr as &$val){
    if(substr($val,0,5) == "delay") $val = "[[" . $val;
}

var_dump($arr);

https://3v4l.org/sIui1

answered Jul 26, 2019 at 12:06

Andreas

24k7 gold badges34 silver badges75 bronze badges

4 Comments

Andreas Over a year ago

@AnantSingh---AlivetoDie Since OPs own regex is hardcoded "delay" then I think we can assume it's not an issue

Death-is-the-real-truth Over a year ago

I am agreed, but still you can add a more generic way too, apart from current answer. There is no harm in that i think :). 3v4l.org/7H1AR And 3v4l.org/npM0p

Aleksa Arsić Over a year ago

This will fail if text with pattern "[[some text]]" is present in text. Thanks for the answer! :)

Andreas Over a year ago

@AleksaArsić always include the fringe cases in your question. How should we know it was even possible to have [[some text]]? It's easily fixed but I see no point in doing so because I'm quite sure there will be a new case coming around the corner.

Collectives™ on Stack Overflow

Parse string with regex and get desired output

3 Answers 3

6 Comments

Comments

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

6 Comments

Comments

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related