Command line(csc-like compiler) parsing

Question

I'm looking for an algorithm/library to parse this command line:

var cmd = "foo /option:value /option2:valueof /etc:.. baa /t:4 etc xd";

to:

foo => { "option", "value" },
       { "option2", "valueof" },
       {"etc", ".."},
baa => {"t", "4"},
etc => {"xd", ""},

I tried in 'pure mode'(C-like) just using if and for. But an solution with regular expression or linq is very appreciated.

my code(not working):

var cmd = "foo /option:value /option2:valueof /etc:.. baa /t:4 etc xd";
            var keys = new Dictionary<string, Dictionary<string,string>>();

            for (int pos = 0, len = cmd.Length; pos < len; pos++)
            {
                StringBuilder buffer = new StringBuilder();
                char c = cmd[pos];

                if (c == '/') 
                {
                    StringBuilder optionName = new StringBuilder();
                    StringBuilder optionValue = new StringBuilder();

                    do
                    {
                        c = cmd[pos++];

                        if (c == ':')
                        {
                            do
                            {
                                c = cmd[pos++];
                                optionValue.Append(c);

                            } while (c != '/' || c != ' ');
                        }
                        else
                        {
                            optionName.Append(c);
                        }

                    } while (c != ' ');

                    keys.Add(buffer.ToString(),
                        new Dictionary<string, string>
                        {
                            {optionName.ToString(), optionValue.ToString()}
                        });
                }
            }

it given an Index was outside the bounds of the array.

Thanks in advance.

Can option values have spaces in them? If so, can you escape them by enclosing them in quotes? Can arguments occur without an option? Can an option occur without a value? And is there a reason that you need to try to parse three commands on one line? Answer to these questions will help us answer your question. — David Merriman, Feb 10 '12 at 18:37

score 2 · Accepted Answer · edited May 23 '17 at 11:48

There are certainly libraries to handle command-line parsing (SO recommends NDesk, but I haven't personally used it).

I'd use string.Split() over iterating character by character through the string.

            var tokenCmd = cmd.Split(' ');
            string currentKey = "";

            foreach (var token in tokenCmd)
            {
                if ((char.IsLetterOrDigit(token[0])) &&
                    (!keys.ContainsKey(currentKey)) ||
                    (keys[currentKey].Any()))
                {
                    currentKey = token;
                    keys.Add(currentKey,
                             new Dictionary<string, string>());
                }
                else
                {
                    var splitToken = new[] { token, "" };

                    if (token.Contains(':'))
                    {
                        splitToken = token
                            .Replace("/", "")
                            .Split(':');
                    }

                    keys[currentKey].Add(splitToken[0],
                                         splitToken[1]);
                }
            }

score 1 · Answer 2 · answered Feb 10 '12 at 18:41

1

change:

} while (c != '/' || c != ' ');

to:

} while (c != '/' && c != ' ');

answered Feb 10 '12 at 18:41

Brandon

983
6
15

xcud · Answer 3 · 2012-02-10T21:12:31.370

If you can retain keep the command lines separate (they're tarred together in your input var) and retain the string[] object type that the args came in on then I think this is the droid you're looking for:

static void Main(string[] args)
{
    Hashtable parsedArgs = new Hashtable();
    args.ToList().ForEach(x => {
        int dpos = x.IndexOf(":");
        if (dpos > -1)
            parsedArgs[x.Substring(1, dpos - 1)] = x.Substring(dpos + 1);
        else
            parsedArgs[x.Substring(1)] = true;
    });
}

Edit: Revised to Linq into a Hashtable as shown here. The Hashtable is much easier to work with than the more elegant-seeming (less code) KeyValuePair<> in my original answer.

Command line(csc-like compiler) parsing

3 Answers3