split string with regular expression by elements

Question

What is a c sharp option to split this string:

"['A','B', ''],['A','D', 'F'],['A','G', 'G']"

into list of strings:

"['A','B', '']"
"['A','D', 'F']"
"['A','G', 'G']"

Answer 1

You're better off writing a simple parser than trying to match balanced text with a regular expression:

var str = "['A','B', ''],['A','D', 'F'],['A','G', 'G']";
var topLevelLists = new List<string>();
var arrStart = -1;
var nesting = 0;
for (int i = 0; i != str.Length; ++i) {
    if (str[i] == '[') {
        if (nesting == 0) {
            arrStart = i;
        }
        ++nesting;
    }
    else if (str[i] == ']') {
        if (nesting <= 0) {
            // Error, ']' without matching '[' at i
            break;
        }
        --nesting;
        if (nesting == 0) {
            topLevelLists.Add(str.Substring(arrStart, i - arrStart + 1));
        }
    }
}
if (nesting > 0) {
    // Error, unmatched '[' at arrStart
}

// topLevelLists => [ "['A','B', '']", "['A','D', 'F']", "['A','G', 'G']" ];

Answer 2

You can use this regex: (?<=\]), to split by comma which ispreceded by ].

The code is:

String input = "['A','B', ''],['A','D', 'F'],['A','G', 'G']";
String pattern = @"(?<=\]),";

var split = Regex.Split(input, pattern);

Answer 3

If your separator is ], (comma comes after the the bracket ]) you can use a tricky way:

var parts = string.Join("]" + char.MaxValue, input
           .Split(new[] {"],"}, StringSplitOptions.None))
           .Split(char.MaxValue);

This approach simply replace all commas that comes after square brackets with a temporary char (char.MaxValue in this case) and then Split the string by that charachter.

Answer 4

The "pure Regex" answer is like this:

        string str = "['A','B', ''],['A','D', 'F'],['A','G', 'G']";
        string[] strs =
            Regex.Matches(str, @"(\[.*?\])")
                 .OfType<Match>()
                 .Select(m => m.Groups[0].Value)
                 .ToArray();

which is more tolerant of different - or even mixed - separators between the bracketed groups, such as comma-space or space rather than just a comma. If your input string is well-defined then this won't be an issue, but I prefer to be able to handle inputs which might come from different sources and might not quite conform.

Answer 5

If I understand, you want to split String with another String. Split String by String is done as described here

For your question - split it by ],