Using a regex with 'or' operator and getting matched groups?

Question

I have some string in a file in the format

rid="deqn1-2"  
rid="deqn3"  
rid="deqn4-5a"  
rid="deqn5b-7"  
rid="deqn7-8"  
rid="deqn9a-10v"  
rid="deqn11a-12c"

I want a regex to match each deqnX-Y where X and Y are either both integers or both combination of integer and alphabet and if there is a match store X and Y in some variables. I tried using the regex (^(\d+)-(\d+)$|^(\d+[a-z])-(\d+[a-z]))$ , but how do I get the values of the matched groups in variables? For a match between two integers the groups would be (I think)

Groups[2].Value 
Groups[3].Value

and for match between two integer and alphabet combo will be

Groups[4].Value
Groups[5].Value

How do I determine which match actually occured and then capture the matching groups accordingly?

won't this `deqn(\w+)-(\w+)` solve your problem? if you want to know which match occurred you can use if else condition — rock321987, Jan 28 '18 at 05:21
@rock321987 `deqn(\w+)-(\w+)` will match `rid="deqn4-5a" ,rid="deqn5b-7"` also, I only want the **digit-digit** matches and get the values of each **digit** or the **digitAlpha-digitAlpha** matches and get each of their value in variables.. — Don_B, Jan 28 '18 at 05:25
if there is digit-digit match or digitAlpha-digitAlpha match, do you want to keep the matched data in same capturing groups? — rock321987, Jan 28 '18 at 05:51
i was thinking to use branch reset i.e. `(?|)` but its not supported in C# — rock321987, Jan 28 '18 at 05:57
right now I can think of only `if else` as an alternative.. maybe someone else can help — rock321987, Jan 28 '18 at 06:04
To determine which match actually occurred, just test which ones are empty... — Nyerguds, Feb 01 '18 at 14:08

rock321987 · Answer 1 · 2018-01-28T06:39:15.753

As branch reset(?|) is not supported in C#, we can use named capturing group with same name like

deqn(?:(?<match1>\d+)-(?<match2>\d+)|(?<match1>\d+\w+)-(?<match2>\d+\w+))\b

regextester demo

C# code

String sample = "deqn1-2";
Regex regex = new Regex("deqn(?:(?<match1>\\d+)-(?<match2>\\d+)|(?<match1>\\d+\\w+)-(?<match2>\\d+\\w+))\\b");

Match match = regex.Match(sample);

if (match.Success) {
    Console.WriteLine(match.Groups["match1"].Value);
    Console.WriteLine(match.Groups["match2"].Value);
}

dotnetfiddle demo

score 0 · Answer 2 · answered Feb 01 '18 at 14:19

You could simply not care. One of the pairs will be empty anyway. So what if you just interpret the result as a combination of both? Just slap them together. First value of the first pair plus first value of the second pair, and second value of the first pair plus second value of the second pair. This always gives the right result.

Regex regex = new Regex("^deqn(?:(\\d+)-(\\d+)|(\\d+[a-z])-(\\d+[a-z]))$");
foreach (String str in listData)
{
    Match match = regex.Match(str);
    if (!match.Success)
        continue;
    String value1 = Groups[1].Value + Groups[3].Value;
    String value2 = Groups[2].Value + Groups[4].Value;
    // process your strings
    // ...
}

Using a regex with 'or' operator and getting matched groups?

2 Answers2