Regex matches but .scan returns nil

Question

As you can see on Rubular the regexp <p( style=".+"){0,1}>.+<\/p> matches the string <p>aasdad</p>.

But, when I do "<p>sdasdasd</p>".scan(/<p( style=".+"){0,1}>.+<\/p>/) I get [[nil]]. Why the matched string is not included in the return value?

Heh just use `?` instead of `{0,1}`. Also, what do you expect ? The group is not matched there and since it's optional the whole expression does match, thus the "matched" group returns nill — HamZa, Jul 30 '13 at 10:30
I also suggest not trying to parse HTML with regexes. Use an HTML parser like Nokogiri. — John Dvorak, Jul 30 '13 at 11:11
@JoeHalfFace most often you don't know the data you'll be getting. For example, your regex will fail if the `p`aragraph has other attributes as well. Even if you know some regex will match what it should, you can never ensure it won't match something it shouldn't, such as `
` in `` or in ``. The latter is not valid HTML but it doesn't mean it won't occur in your data. Also, some people like not to escape HTML in `textarea`s. Also, `
` is fun. — John Dvorak, Jul 30 '13 at 11:20
@JanDvorak, this is to refactor some of redactor.js output server side, actually, it has very limited nubmer of options — Joe Half Face, Jul 30 '13 at 11:29
@JoeHalfFace I'd still go for Nokogiri. You never know what version changes bring. Scraping HTML is always prone to failing due to HTML changes. Using DOM reduces the risk somewhat. Using DOM with good selectors reduces the risk significantly. — John Dvorak, Jul 30 '13 at 11:35

score 6 · Accepted Answer · answered Jul 30 '13 at 10:33

6

That's the way scan works. From the Ruby documentation for scan:

If the pattern contains groups, each individual result is itself an array containing one entry per group.

Since the optional group ( style=".+") doesn't match you get only a nil in the result. You can use (?: for a non-capturing group:

"<p>sdasdasd</p>".scan(/<p(?: style=".+"){0,1}>.+<\/p>/)
# => ["<p>sdasdasd</p>"]

answered Jul 30 '13 at 10:33

toro2k

@Joe Half Face - apparently you hadn't, or else you wouldn't have needed to ask. – pguardiario Jul 30 '13 at 12:34
@pguardiario, after I asked question there was sudden idea) – Joe Half Face Jul 30 '13 at 12:54
@Joe Half Face - maybe you should have considered the problem longer before posting then. You just earned a downvote from me. – pguardiario Jul 30 '13 at 12:56
1

@pguardiario, I don't know, man, may be you can manage the time you understand things in seconds scale or forecast future, not everyone can. – Joe Half Face Jul 30 '13 at 20:04

score 0 · Answer 2 · answered Jul 30 '13 at 10:38

0

You could also try with .match

"<p>sdasdasd</p>".match(/<p( style=".+"){0,1}>.+<\/p>/)
# => <p>sdasdasd</p>

answered Jul 30 '13 at 10:38

Nicolas Duval

Not really, `match` returns a `MatchData` object, not a string, you should do `"...".match(/.../)[0]` to get the matched string. – toro2k Jul 30 '13 at 10:40
1

@toro2k, or just use String#[] like most people. – pguardiario Jul 30 '13 at 12:36
`MatchData` also exposes a `#captures` instance method which will return any captured MatchData. – lightyrs Nov 18 '14 at 22:53

2 Answers2