Here is a cool little method that I modifed from a VB example. Does exactly
what I wanted. Can split on any delimiter or multiple delimiters and can
quote using any pair or chars. Very cool. Have not tested all possible
failures, etc, but appears to work well. Some clever (and generous) pattern
person may want to modify this to allow an *array of quote pairs, so you
could quote on "one two" or {one two} or (one two) in the same call. If you
do, please post update. Cheers!
==
/// <summary>
/// Split a string, dealing correctly with quoted items.
/// The quotes parm is the character pair used to quote strings
/// (default is "", the double quote).
/// You can also use a character pair (eg "{}") if the opening
/// and closing quotes are different.
///
/// For example, you can split the following string:
/// string[] fields = SplitQuoted("[one,two],three,[four,five]", , "[]")
/// into 3 items, because commas inside [] are not taken into account.
/// </summary>
/// <remarks>
/// Multiple seperators are ignored, so splitting "a,,b" using a comma as
/// the seperator will return two fields, not three. To get this behavior,
/// you could use ", " (comma and space) as seperators and default quotes.
/// Then set the string to something like ' a, "", b ' to get the empty
field.
/// You could also use comma as *only seperator and put a space to get a
space field
/// like 'a, ,b'.
/// </remarks>
/// <param name="text">string to split</param>
/// <param name="seperator">The seperator char(s) as string.</param>
/// <param name="quotes">The char pair used to quote a string.</param>
/// <returns>string[]</returns>
private string[] SplitQuoted(string text, string seperators, string quotes)
{
// Default seperators is a space and tab (e.g. " \t").
// All seperators not inside quote pair are ignored.
// Default quotes pair is two double quotes ( e.g. '""' ).
if ( text == null )
throw new ArgumentNullException("text", "text is null.");
if ( seperators == null || seperators.Length < 1 )
seperators = " \t";
if ( quotes == null || quotes.Length < 1 )
quotes = "\"\"";
ArrayList res = new ArrayList();
// Get the open and close chars, escape them for use in regular
expressions.
string openChar = Regex.Escape(quotes[0].ToString());
string closeChar = Regex.Escape(quotes[quotes.Length - 1].ToString());
// Build the pattern that searches for both quoted and unquoted elements
// notice that the quoted element is defined by group #2
// and the unquoted element is defined by group #3.
string pattern = @"\s*(" + openChar + "([^" + closeChar + "]*)" +
closeChar + @"|([^" + seperators + @"]+))\s*";
// Search the string.
foreach ( System.Text.RegularExpressions.Match m in
System.Text.RegularExpressions.Regex.Matches(text, pattern) )
{
string g3 = m.Groups[3].Value;
if ( g3 != null && g3.Length > 0 )
res.Add(g3);
else
{
// get the quoted string, but without the quotes.
res.Add(m.Groups[2].Value);
}
}
return (string[])res.ToArray(typeof(string));
}
ne