Why is it matching two Properties - regex

Question

I have a regex where I want to get Properties which contains %%text%%, but my regex is capturing more than that

My regex: (<Properties>).+?%%.+?%%.+?(<\/Properties>)

It maches:

"<Properties>
<Property>TEXT</Property>
</Properties>
<Properties>
<Property >%%TEXT%%</Property>
</Properties>"

But I want him to match only:

"<Properties>
<Property >%%TEXT%%</Property>
</Properties>"

What am I doing wrong?

You can't parse [X]HTML with regex. Because HTML can't be parsed by regex https://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags — ButchMonkey, Jan 23 '20 at 15:00
Could you not just use `%%.*%%<\/Property><\/Properties>` ? Assuming you have only one "property" in each "properties"? [Test here](https://regex101.com/r/cUy5YP/2) — ButchMonkey, Jan 23 '20 at 15:08
I have big XML String, where in some Properties I have variables with %%something%%, and I need to find all Properties who containains it and loop them trough and replace %%something%% with actual value. I can take all properties, and loop them trough and find this whom has this variable, but that ir overhead to my server — Jānis Seržants, Jan 23 '20 at 15:15
How about using [XPath](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Introduction_to_using_XPath_in_JavaScript) instead? — Andrew Morton, Jan 23 '20 at 15:30

Toto · Accepted Answer · 2020-01-23T16:12:25.910

3

Use tempered greedy token instead of .:

<Properties>(?:(?!<\/Properties>)[^])*%%(.+?)%%(?:(?!<\/Properties>)[^])*<\/Properties>

This part (?:(?!<\/Properties>)[^]) makes sure we haven't </properties> before the wanted text.

[^] stands for any character including newlines.

Demo

edited Jan 23 '20 at 16:12

answered Jan 23 '20 at 15:14

Toto

89,455
62
89
125

Take a look: https://regex101.com/r/hN1yId/4 It still has issues :( – Jānis Seržants Jan 23 '20 at 15:21
1

@JānisSeržants: It had, I've done a correction, revisit the link provided in demo. – Toto Jan 23 '20 at 15:23
I was learning from w3schools (https://www.w3schools.com/jsref/jsref_obj_regexp.asp) about regex, but now I see there aren't full list of regex options, where can I read more about for example `?:` ? – Jānis Seržants Jan 23 '20 at 15:45
@JānisSeržants https://www.regular-expressions.info/brackets.html – 3limin4t0r Jan 23 '20 at 15:49
@JānisSeržants: [Tempered greedy token](http://www.rexegg.com/regex-quantifiers.html#tempered_greed) – Toto Jan 23 '20 at 15:56

3limin4t0r · Answer 2 · 2020-01-23T15:35:01.387

Let's break down the regex vs the actual match so you can see why it matches:

(<Properties>).+?%%.+?%%.+?(<\/Properties>)

(<Properties>) matches the first <Properties>.
.+? matches one or more characters until it encounters %%, thus matching <Property>TEXT</Property><Properties><Property >.
%% matches %%.
.+? matches one or more characters until it encounters %%, thus matching TEXT.
%% matches %%.
.+? matches one or more characters until it encounters </Properties> thus matching </Property.
(<\/Properties>) matches </Properties>.

Instead you want to make your regex more explicit:

(?:[^<%]|%(?!%)|<(?!\/Properties>))

The above will match one character that is not < or %, if it is one of those two it will only match % if not followed by another % and it will only match < if not followed by /Properties>. This should be used as replacement for your .. Resulting in:

(<Properties>)(?:[^<%]|%(?!%)|<(?!\/Properties>))+%%(?:[^<%]|%(?!%)|<(?!\/Properties>))+%%(?:[^<%]|%(?!%)|<(?!\/Properties>))+(<\/Properties>)

Since the regex is more explicit I can remove the lazy ? quantifier modifier safely.

Why is it matching two Properties - regex

2 Answers2