Is there an equivalent to 'sscanf()' in .NET?

Question

The .NET Framework gives us the Format method:

string s = string.Format("This {0} very {1}.", "is", "funny");
// s is now: "This is very funny."

I would like an "Unformat" function, something like:

object[] params = string.Unformat("This {0} very {1}.", "This is very funny.");
// params is now: ["is", "funny"]

I know something similar exists in the ANSI-C library (printf vs scanf).

The question: is there something similiar in C#?

Update: Capturing groups with regular expressions are not the solution I need. They are also one way. I'm looking for a system that can work both ways in a single format. It's OK to give up some functionality (like types and formatting info).

Answer 1

You could do string[] parts = string.Split(' '), and then extract by the index position parts[1] and parts [3] in your example.

Answer 2

You'll need to write a regular expression to get this type of functionality.

Answer 3

Regex with grouping?

/This (.*?) very (.*?)./

Answer 4

There's no such method, probably because of problems resolving ambiguities:

string.Unformat("This {0} very {1}.", "This is very very funny.")
// are the parameters equal to "is" and "very funny", or "is very" and "funny"?

Regular expression capturing groups are made for this problem; you may want to look into them.

Answer 5

Yep. These are called "regular expressions". The one that will do the thing is

This (?<M0>.+) very (?<M1>.+)\.

Answer 6

Using regular expressions is your best bet. Investigate the Regex class.

Answer 7

@mquander: Actualy, PHP solves it even different:

$s = "This is very very funny.";
$fmt = "This %s very %s.";
sscanf($s, $fmt, $one, $two);
echo "<div>one: [$one], two: [$two]</div>\n";
//echo's: "one: [is], two: [very]"

But maybe your regular expression remark can help me. I just need to rewrite "This {0} very {1}." to something like: new Regex(@"^This (.*) very (.*)\.$"). This should be done programmatical, so I can use one format string on the public class interface.

BTW: I've already have a parser to find the parameters: see the Named Format Redux blog entry by Phil Haack (and yes, I also want named paramters to work both ways).

Answer 8

I came across the same problem, i belive that there is a elegante solution using REGEX... but a came up with function in C# to "UnFormat" that works quite well. Sorry about the lack of comments.

    /// <summary>
    /// Unformats a string using the original formating string. 
    /// 
    /// Tested Situations:
    ///    UnFormat("<nobr alt=\"1\">1<nobr>", "<nobr alt=\"{0}\">{0}<nobr>") : "1"
    ///    UnFormat("<b>2</b>", "<b>{0}</b>") : "2"
    ///    UnFormat("3<br/>", "{0}<br/>") : "3"
    ///    UnFormat("<br/>4", "<br/>{0}") : "4"
    ///    UnFormat("5", "") : "5"
    ///    UnFormat("<nobr>6<nobr>", "<nobr>{0}<nobr>") : "6"
    ///    UnFormat("<nobr>2009-10-02<nobr>", "<nobr>{0:yyyy-MM-dd}<nobr>") : "2009-10-02"
    ///    UnFormat("<nobr><nobr>", "<nobr>{0}<nobr>") : ""
    ///    UnFormat("bla", "<nobr>{0}<nobr>") : "bla"
    /// </summary>
    /// <param name="original"></param>
    /// <param name="formatString"></param>
    /// <returns>If an "unformat" is not possible the original string is returned.</returns>
    private Dictionary<int,string> UnFormat(string original, string formatString)
    {
       Dictionary<int, string> returnList = new Dictionary<int, string>();

       try{
          int index = -1;

          // Decomposes Format String
          List<string> formatDecomposed = new List<string> (formatString.Split('{'));
          for(int i = formatDecomposed.Count - 1; i >= 0; i--)
          {
             index = formatDecomposed[i].IndexOf('}') + 1;

             if (index > 0 && (formatDecomposed[i].Length - index) > 0)
             {
                formatDecomposed.Insert(i + 1, formatDecomposed[i].Substring(index, formatDecomposed[i].Length - index));
                formatDecomposed[i] = formatDecomposed[i].Substring(0, index);
             }
             else
                //Finished
                break;
          }

          // Finds and indexes format parameters
          index = 0;
          for (int i = 0; i < formatDecomposed.Count; i++)
          {
             if (formatDecomposed[i].IndexOf('}') < 0)
             {
                index += formatDecomposed[i].Length;
             }
             else
             {
                // Parameter Index
                int parameterIndex;
                if (formatDecomposed[i].IndexOf(':')< 0)
                   parameterIndex = Convert.ToInt16(formatDecomposed[i].Substring(0, formatDecomposed[i].IndexOf('}')));
                else
                   parameterIndex = Convert.ToInt16(formatDecomposed[i].Substring(0, formatDecomposed[i].IndexOf(':')));

                // Parameter Value
                if (returnList.ContainsKey(parameterIndex) == false)
                {
                   string parameterValue;

                   if (formatDecomposed.Count > i + 1)
                      if (original.Length > index)
                         parameterValue = original.Substring(index, original.IndexOf(formatDecomposed[i + 1], index) - index);
                      else
                         // Original String not valid
                         break;
                else
                   parameterValue = original.Substring(index, original.Length - index);

                returnList.Add(parameterIndex, parameterValue);
                index += parameterValue.Length;
             }
             else
                index += returnList[parameterIndex].Length;

             }
          }

          // Fail Safe #1
          if (returnList.Count == 0) returnList.Add(0, original);
       } 
       catch
       {
          // Fail Safe #2
          returnList = new Dictionary<int, string>();
          returnList.Add(0, original);
       }

       return returnList;
    }