Regex doesn't match as expected

Question

I'm trying to write a regex to match the following text:

one of letters f, s or t
multiple times:
- at least one digit
- one of letters f, s or t

This is the regex I've come up with: (f|s|t)(\d+(f|s|t))*, but it doesn't seem to work.

Input: f20s30t

Expected matches:

f
20s
30t

Actual matches:

f
30t
t

Why is my regex expression wrong?

Edit: I'm using this method to split the input.

By "matches", you meant captures, right? It is not quite clear what you are doing. You may actually just get multiple matches with \d*[fst]. In .NET, you may grab all the captures by accessing the capture collection related to a certain group. Is it a whole string or part of a longer text? — Wiktor Stribiżew
– Wiktor Stribiżew, Commented May 17, 2018 at 15:20
@WiktorStribiżew I'm using Regex split in C#, but I think that's kind of irrelevant to the question. You can test the regex yourself in regex101.com — Tirafesi
– Tirafesi, Commented May 17, 2018 at 15:24
Oh, that is very relevant because you should not test the .NET regex patterns at regex101 as it does not support .NET regex. So, is that a whole string or part of a longer text? — Wiktor Stribiżew
– Wiktor Stribiżew, Commented May 17, 2018 at 15:25
I think you are rather interested in match groups. Are you? You've given only one example - not sufficient when you're asking strangers about a regex. Give a couple of more positive examples as well as negative examples. — Aditya
– Aditya, Commented May 17, 2018 at 15:26

Wiktor Stribiżew · Accepted Answer · 2018-05-17 15:31:39Z

Since you are using C#, you may match a string with a pattern and access the capture collection of each repeated capturing group:

var s = "f20s30t"; var m = Regex.Match(s, @"^([fst])(\d+[fst])*$"); if (m.Success) { Console.WriteLine(m.Groups[1].Value); foreach (var g in m.Groups[2].Captures.Cast<Capture>().Select(t => t.Value)) Console.WriteLine(g); }

The advantage of this approach is that it also validates the string, and you won't get results for a string like TEXT random f20s30t here.....

See the C# demo, output

f 20s 30t

Here is the regex demo:

Details

^ - start of the string
([fst]) - Capturing group 1: f, s or t
(\d+[fst])* - 0 or more repetitions (captured into Group 2 with each value saved in the group stack) of:
- \d+ - 1+ digits
- [fst] - f, s or t
$ - end of string.

James Whiteley · Accepted Answer · 2018-05-17 15:35:05Z

Going by your example code:

^([fst]|[0-9]+[fst])$

^: start of string
[fst]: one of f, s or t
|: OR
[0-9]+: one or more digits
[fst]: one of f, s or t
$: end of string

In C#:

using System; using System.Text.RegularExpressions; public class Program { public static void Main() { string input = "f20s30t"; string pattern = "([fst]|[0-9]+[fst])"; string[] substrings = Regex.Split(input, pattern); foreach (string match in substrings) { Console.WriteLine("'{0}'", match); } } } // //f // //20s // //30t

.NET Fiddle

Regex101

revo · Accepted Answer · 2018-05-17 15:44:59Z

In regex world (f|s|t) is equal to [fst] but has a few downsides in comparison. So use latter (a character class) in place of former. Using split method you could split on this:

(?<=[fts])(?=\d+[fts])

Above regex matches a position where a character from [fts] meets an occurrence of digits following a character from [fts]. This (?<=...) is a positive look-behind and (?=...) constructs a positive look-ahead.

RegEx live demo

Code (see demo here):

string input = "f20s30t"; string pattern = @"(?<=[fts])(?=\d+[fts])"; string[] substrings = Regex.Split(input, pattern); foreach (string match in substrings) { Console.WriteLine("{0}", match); }

Output:

f 20s 30t

Collectives™ on Stack Overflow

Regex doesn't match as expected

3 Answers 3

Comments

Comments

Comments

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

Comments

Related