Regex match multiple lines before and after word delimited by start and end words

Question

I am wanting to search for {{ upc }} and start the capture not from the <div immediately ahead of the match but the 2nd <div ahead of the match i.e. <div class="form-group"> and capture not up to the first </div> after the match but the 2nd i.e closing </div> or up to the start of the next <div class="form-group"> (depending on how you look at it)

Here is the sample HTML/Twig template I am wanting to search and replace.

<div class="form-group">
    <label class="col-sm-2 control-label" for="input-sku"><span data-toggle="tooltip" title="{{ help_sku }}">{{ entry_sku }}</span></label>
    <div class="col-sm-10">
        <input type="text" name="sku" value="{{ sku }}" placeholder="{{ entry_sku }}" id="input-sku" class="form-control"/>
    </div>
</div>
<div class="form-group">
     <label class="col-sm-2 control-label" for="input-upc"><span data-toggle="tooltip" title="{{ help_upc }}">{{ entry_upc }}</span></label>
     <div class="col-sm-10">
         <input type="text" name="upc" value="{{ upc }}" placeholder="{{ entry_upc }}" id="input-upc" class="form-control"/>
     </div>
</div>
<div class="form-group">
     <label class="col-sm-2 control-label" for="input-ean"><span data-toggle="tooltip" title="{{ help_ean }}">{{ entry_ean }}</span></label>
     <div class="col-sm-10">
         <input type="text" name="ean" value="{{ ean }}" placeholder="{{ entry_ean }}" id="input-ean" class="form-control"/>
     </div>
</div>

The expected regex match is as follows:

<div class="form-group">
     <label class="col-sm-2 control-label" for="input-upc"><span data-toggle="tooltip" title="{{ help_upc }}">{{ entry_upc }}</span></label>
     <div class="col-sm-10">
         <input type="text" name="upc" value="{{ upc }}" placeholder="{{ entry_upc }}" id="input-upc" class="form-control"/>
     </div>
</div>

All help appreciated. Thank you.

@samthegolden I'm starting with `/^.*(?={{ sku }}).*/gm` on regex101 although that is probably too vague for a starting point. I will keep fumbling and if I solve post it here but I suspect a regex master will beat me to it! — Trent Renshaw, Apr 20 '20 at 08:35

samthegolden · Answer 1 · 2020-04-20T12:45:23.833

0

You need to parse the div's you want and then absorb everything inside them and exclude the rest.

[\w\W] means match words and non-words. It matches newline characters for instance, which * does not.

[\w\W]*(<div[\w\W]*?<div[\w\W]*?{{ sku }}[\w\W]*?<\/div>[\w\W]*?<\/div>)[\w\W]*

edited Apr 20 '20 at 12:45

answered Apr 20 '20 at 08:59

samthegolden

1,366
1
10
26

thanks but it's far too greedy with too many steps. The HTML/Twig code is just a small part of the whole document. Unfortunately this regex matches everything before and after `{{ sku }}` In the end I decided to start capturing from left to right starting with the opening `
` matching `{{ sku }}` (non-capturing) in the middle up until the next starting `
` as follows: `(
.*(?:{{ sku }}).*)(?:
)`
– Trent Renshaw Apr 20 '20 at 09:13
You said you wanted to capture 2 div's ahead of the chosen word... – samthegolden Apr 20 '20 at 09:26
And btw that does not match your example: https://regex101.com/r/QkK8LY/1 – samthegolden Apr 20 '20 at 09:28
You are using `/gm` modifier not `/s` for dot to match newline characters example: https://regex101.com/r/yFPT9o/1 .. The regex still is not the solution as it will start match from the first `
` through to `{{ sku }}`. I think this will be a task for look ahead and look behind.
– Trent Renshaw Apr 20 '20 at 10:32
How does my regex not fulfil your requirements? If you explain better I may improve it – samthegolden Apr 20 '20 at 10:40
see your regex and test string in v2 https://regex101.com/r/QkK8LY/2 .. I have expanded the test string to give more context / clarity – Trent Renshaw Apr 20 '20 at 11:42
Check my edit @TrentRenshaw. I was not matching spaces and variations... This way you don't need to specify an initial pattern and don't need to use lookaheads, but use non-greedy capture. – samthegolden Apr 20 '20 at 12:45

score 0 · Accepted Answer · answered Apr 27 '20 at 15:06

One thing you can try is to use a negative lookahead to filter out the things you do not wish to be included in your match. For instance, matching a <div, followed by anything and then another <div, can match things like <div></div><div>.

Instead, what you can say is to match <div, followed by anything - as long as it is not </div> - and then another <div.

<div    (?:(?!</div>).)*    <div

Then, you can insert that same subpattern anywhere in your expression where you'd normally write .*. In this particular case, you can repeat that to make sure you're not hitting a closing div before the UPC and then continue with the {{ UPC }} portion.

<div(?:(?!</div>).)*<div    (?:(?!</div>).)*    {{ upc }}    .*?</div>\s*</div>

Here is a demo

Perfect and much appreciated. Thank you! – Trent Renshaw Apr 28 '20 at 06:09 — Trent Renshaw, Apr 28 '20 at 06:09

Regex match multiple lines before and after word delimited by start and end words

2 Answers2