C# Regular Expression Capturing Empty String

Question

I'm trying to create a simple regular expression in C# to split a string into tokens. The problem I'm running into is that the pattern I'm using captures an empty string, which throws off my expected results. What can I do to change my regular expression so it doesn't capture an empty string?

var input = "ID=123&User=JohnDoe";
var pattern = "(?:id=)|(?:&user=)";
var tokens = Regex.Split(input, pattern, RegexOptions.IgnoreCase);

// Expected Results
// tokens[0] == "123"
// tokens[1] == "JohnDoe"

// Actual Results
// tokens[0] == ""
// tokens[1] == "123"
// tokens[2] == "JohnDoe"

See [Easiest way to parse “querystring” formatted data](https://stackoverflow.com/questions/11956948/easiest-way-to-parse-querystring-formatted-data). — Wiktor Stribiżew, Sep 22 '17 at 20:39
link doesn't answer OP - it relies on a System.Net.Http static method that doesn't exist in .NET core — Josh E, Sep 22 '17 at 20:41
@WiktorStribiżew definitely points to a better approach. Your query string would fail on semantically identical input `"User=JohnDoe&ID=123"` because of your check for an `&` in the regex. It's best not to reinvent the wheel on this one. — Corey Ogburn, Sep 22 '17 at 20:43
There several ways to fix this. 1) [Remove empty items](https://stackoverflow.com/questions/4912365/c-sharp-regex-split-removing-empty-results), 2) Use `(?i)(?<=id=)[^&]+` to get id and `(?i)(?<=user=)[^&]+` to get user name, 3) etc. — Wiktor Stribiżew, Sep 22 '17 at 20:46

Josh E · Accepted Answer · 2017-09-22T20:52:57.207

While the comments to your OP regarding using a different approach may have merit, they don't address your specific question regarding the RegEx behavior.

I think that the reason though you're getting the regex behavior has to do with an implicit capture group (ed: or it could just be limiting the capture behavior of the first group is sufficient), but I haven't made it to the top level of the RegEx hierarchy of understanding.

Edit:

Working RegEx for the given test case:

(?>id=)|(?:&user=)

If none of this is to your liking, you could always tack a predicate to the tokens list:

tokens.Where(x => !string.IsNullOrWhiteSpace(x))

score 0 · Answer 2 · answered Sep 22 '17 at 20:47

I don't think you can solve this problem with Regex.Split to be honest. One brute force way to do this is to remove every "":

var input = "ID=123&User=JohnDoe";
var pattern = "(?:id=)|(?:&user=)";
var tokens = Regex.Split(input, pattern, RegexOptions.IgnoreCase).Where(x => x != "");

I think you should use regex that actually captures the tokens in groups.

var input = "ID=123&User=JohnDoe";
var pattern = "id=(.+)&user=(.+)";
var match = Regex.Match(input, pattern, RegexOptions
    .IgnoreCase);
match.Groups[1] // 123

C# Regular Expression Capturing Empty String

2 Answers2