Regular Expression lazy modifier matches too much

Question

The following regular expression is jumping [url] tags...

Regular Expression (generic regular expression)

(?:\[url.*?\])(.*?youtu.*?)(?:\[\/url\])

String:

[url]blahyoutubeblah[/url] heyya [url]blahblah[/url]    [url]www.youtube.com/blah[/url]

Help!!

also you should say what flavour of regex you are using..... I presume js — Keith Nicholas, Sep 09 '18 at 23:14
it fails because you say match any character until youtu, then look for /url — Keith Nicholas, Sep 09 '18 at 23:16
@KeithNicholas yes... js or PHP.... thanks but I thought the lazy modifier ***.*?*** would fix that so it would get minimum length string. ?? — zdanman, Sep 09 '18 at 23:18

score 2 · Accepted Answer · answered Sep 09 '18 at 23:24

Your captured group requires youtu inside, so the substring

[url]blahblah[/url]    [url]www.youtube.com/blah[/url]

matches, because it starts with [url], includes youtu, and ends with [/url].

Simply using a negated character set, excluding [, probably isn't enough, because that wouldn't allow for nested tags to match, such as an input of

[url]foobar youtube[b]BOLD TEXT[/b][/url]

You might require negative lookahead for [/url] right before each repeated character:

(?:(?!\[\/url\]).)*

Also, make sure that whatever comes after the [url does not contain ]s before coming to the true ], with:

\[url[^]]*\]

In full:

\[url[^]]*\]((?:(?!\[\/url\]).)*youtu(?:(?!\[\/url\]).)*)\[\/url\]

There's no need to make the quantifiers lazy anymore, because of the negative lookahead.

Demo:

https://regex101.com/r/hSAJEp/1

you're awesome. why require a negative lookahead for [/url]? before repeated characters? — zdanman, Sep 09 '18 at 23:35
Because otherwise, the substring at the top of the answer (`[url]blahblah[/url] [url]www.youtube.com/blah[/url]`) will match. You don't want any `[/url]`s to be in between the first `[url]` and the final `[/url]`. — CertainPerformance, Sep 09 '18 at 23:41

score 1 · Answer 2 · edited Sep 09 '18 at 23:23

1

You are matching .* which means it will match url, up until youtu, then find /url

A simple workaround could be something like which means it won't match a opening [ bracket before finding youtu

(?:\[url.*?\])([^\[]*?youtu.*?)(?:\[\/url\])

edited Sep 09 '18 at 23:23

Paolo

21,270
6
38
69

answered Sep 09 '18 at 23:20

Keith Nicholas

43,549
15
93
156

Qwertiy · Answer 3 · 2018-09-09T23:37:03.203

1

It's lazy, but it still will match if it can - it won't be moving left border if match is possible. There are other things to do that. One of them is just to prevent unwanted match by regex itself - just use

(?:\[url[^\]]*?\])([^\[]*?youtu.*?)(?:\[\/url\])

edited Sep 09 '18 at 23:37

answered Sep 09 '18 at 23:22

Qwertiy

19,681
15
61
128

you're awesome. this site is awesome. thanks! – zdanman Sep 09 '18 at 23:35
1

@zdanman, if it solves your problem, accept an answer by clicking a check on the left of it. – Qwertiy Sep 09 '18 at 23:37
THANKS Qwertiy I had to end up using the one by CertainPerformance because it allows other tags ( such as **[B][/B]**, **[U][/U]** ) to be included in the group result - and specifically targets the **[URL][/URL]** tag to be excluded. I so very appreciate your help though! Truly! – zdanman Sep 10 '18 at 00:35

score 1 · Answer 4 · answered Sep 09 '18 at 23:30

1

The problem was that there is youtu you had in your regex but there was blahblah between url to be matched, making it generic so

(?:\[url.*?\])(.*?)(?:\[\/url\])

answered Sep 09 '18 at 23:30

The Scientific Method

2,374
2
14
25

2

I think he wants to find url tags with `youtu` inside, but you'll find any. – Qwertiy Sep 09 '18 at 23:35
1

he just said jumping `url` – The Scientific Method Sep 09 '18 at 23:35

Regular Expression lazy modifier matches too much

4 Answers4