regex for optional group

Question

I'd like to parse the following sample string

foo :6

into two groups: Text and Number. The number group should be populated only if the character ":" precedes the number itself.

so:

foo 6 -> Text = "foo 6"
foo :6 -> Text = "foo", Number = "6"

The best I could come up with so far is

(?<Text>.+)(?=:(?<Number>\d+)h?)?

but that doesn't work because the first group greedily expands to the whole string.

Any suggestions?

BlackBear · Answer 1 · 2014-09-12T12:58:54.323

5

If you really want to use a regex you can write quite a simple one, without lookarounds:

(?<Text>[^:]+):?(?<Number>\d*)

In my opinion, regexes should be as simple as possible; if you do not want spaces around the Text group I suggest you use match.Groups["Text"].Value.Strip().

Note that if you are parsing a multiline string this pattern will not work because, as @OscarHermosilla mentioned below, [?:]+ will also match newlines. The fix is simple though, change it with [^:\n]

edited Sep 12 '14 at 12:58

answered Sep 12 '14 at 10:35

BlackBear

20,590
9
41
75

@Unihedron I guess it depends on OP's intentions – BlackBear Sep 12 '14 at 10:46
See test case: `foo :6 -> Text = "foo", Number = "6"` – Unihedron Sep 12 '14 at 10:46
I agree with you regex should be simple and this will do the job. But it should be applied individually to every string to be parsed. But it won't work in a text with several lines becuase [^:]+ will also match new-line. – Oscar Hermosilla Sep 12 '14 at 12:53
@OscarHermosilla you are correct, indeed, I posted this because the question seems to suggest that individual strings will be used. I will add this to my answer – BlackBear Sep 12 '14 at 12:57

Avinash Raj · Accepted Answer · 2014-09-12T12:08:19.710

You don't need any seperate function for stripping the trailing whitespaces

The below regex would capture all the characters into the named group Text except :\d+(ie; : followed by one or more numbers). If it finds a colon followed by numbers, then it starts capturing the number into the named group Number

^(?<Text>(?:(?!:\d+).)+(?=$|\s+:(?<Number>\d+)$))

DEMO

String input = "foo 6";
String input1 = "foo :6";
Regex rgx = new Regex(@"^(?<Text>(?:(?!:\d+).)+(?=$|\s+:(?<Number>\d+)$))");

foreach (Match m in rgx.Matches(input))
{
Console.WriteLine(m.Groups["Text"].Value);
}
foreach (Match m in rgx.Matches(input1))
{
Console.WriteLine(m.Groups["Text"].Value);
Console.WriteLine(m.Groups["Number"].Value);
}

Output:

foo 6
foo
6

IDEONE

Seems like the best match for what I want to do. Thanks. – Chris Sep 12 '14 at 14:01 — Chris, Sep 12 '14 at 14:01

score 1 · Answer 3 · edited May 23 '17 at 11:49

1

You can repeat the group name text with an alternation. This way:

(?<Text>.+)\s+:(?<Number>\d)|(?<Text>.+)

DEMO

Based on the idea behind this post: Regex Pattern to Match, Excluding when... / Except between

edited May 23 '17 at 11:49

Community

1
1

answered Sep 12 '14 at 10:43

Oscar Hermosilla

421
5
20

It could be easily fixed if it is needed – Oscar Hermosilla Sep 12 '14 at 10:49
1

For example: `(?.+)\s+:(?\d)|(?.+)` – Oscar Hermosilla Sep 12 '14 at 10:52
@Unihedron Not really, at least I tried it and it doesn't seem to work http://regex101.com/r/vJ4hI8/1 – Oscar Hermosilla Sep 12 '14 at 11:00

score 0 · Answer 4 · answered Sep 12 '14 at 10:29

0

You can simply use split instead of regex:

"foo :6".Split(':');

answered Sep 12 '14 at 10:29

Rahul Tripathi

152,732
28
233
299

score 0 · Answer 5 · answered Sep 12 '14 at 10:39

0

You can try like:

(\D+)(?:\:(\d+))

or do a Regex.Split using this pattern:

(\s*\:\s*)

answered Sep 12 '14 at 10:39

NeverHopeless

10,503
4
33
53

regex for optional group

5 Answers5