Keep (and associate) delimiters when splitting a string in C#

Question

In spliting a string I want to generate a sequence of token-delimiter pairs. Thus, with , and ; as my delimiters, I want " a , b;" to produce new int[][]{{" a ",","},{" b",";"},{"",""}}. The last entry indicates that the string ends with a delimiter. Of course, two consecutive delimiters are to be separated by and empty token.

Answer 1

Neither String.Split nor Regex.Split allow such an association - the result is always a sequence of strings. Even when also capturing the split token in the sequence as so the delimiter will be mixed-in.

However this task can be accomplished easily with Regex.Matches (or Match/NextMatch). The trick is to use the \G anchor (see Anchors in Regular Expressions) such that the matching is incremental and resumes from the previous match.

var input = @" a , b;whatever";

// The \G anchor ensures the next match begins where the last ended.
// Then non-greedily (as in don't eat the separators) try to find a value.
// Finally match a separator.
var matches = Regex.Matches(input, @"\G(.*?)([,;])")
    .OfType<Match>();

// All the matches, deal with pairs as appropriate - here I simply group
// them into strings, but build a List of Pairs or whatnot.
var res = matches
    .Select(m => "{" + m.Groups[1].Value + "|" + m.Groups[2].Value + "}");
// res -> Enumerable with "{ a |,}", "{ b|;}" 

String trailing;
var lastMatch = matches.LastOrDefault();
if (lastMatch != null) {
    trailing = input.Substring(lastMatch.Index + lastMatch.Length);
    // If the separator was at the end, trailing is an empty string
} else {
    // No matches, the entire input is trailing.
    trailing = input;
}

// trailing -> "whatever"

Have fun filling in the details (and fixing any issues) as required. For tidiness, modify this code as appropriate and put it inside a method.