Question

I’m trying to check a string for key words, and if the word exists, get the value from the dictionary. The problem exists when the key word is a multi-word phrase.

So I have a dictionary:

Dictionary<string, string> d = new Dictionary<string, string>(StringComparer.InvariantCultureIgnoreCase);

d.Add("keyword1", "D2");
d.Add("keyword2", "D3");
d.Add("keyword3", "D4");
d.Add("keyword4", "D4");
d.Add("keyword5", "D5");
d.Add("key word six", "D6"); 

And I have a string, which may look like the following but will be a random sentence:

string errormessage = "This is an error regarding Key Word Six";

I’m currently using the following to check the errormessage and see if any words appear in the dictionary:

string code = null;
string theDcode = null;

foreach (string word in errormessage.Split(' '))
{
    if (d.TryGetValue(word, out theDcode))
    {
        code = theDcode;
    }
}

The problem is that I can’t search for the string “Key Word Six” since I’m reading the string word by word and the foreach loop sees each word separately. This works great for single word key words. How can I handle checking for a multi word key word?

Answer 1

You could iterate the items in the dictionary instead and check the string for matches.

string errormessage = "This is an error regarding Key Word Six";
var d = new Dictionary<string, string>(StringComparer.CurrentCultureIgnoreCase);

d.Add("keyword1", "D2");
d.Add("keyword2", "D3");
d.Add("keyword3", "D4");
d.Add("keyword4", "D4");
d.Add("keyword5", "D5");
d.Add("key word six", "D6"); 

string code = null;
foreach (var item in d)
{
    var i = errormessage.IndexOf(item.Key, StringComparison.CurrentCultureIgnoreCase);
    if(i >= 0)
        code = item.Value;
}

This would however also give you matches inside of words. testkeyword1test for example.

EDIT
For possible better performance (untested) you could use a regular expression.

string code;
var reg = new Regex(GetPatternString(d.Select (x => x.Key)), RegexOptions.IgnoreCase);
foreach (var match in reg.Matches(errormessage))
{
    code = d[match.ToString()];
}

And helper function

private static string GetPatternString(IEnumerable<string> values)
{
    var sb = new StringBuilder();
    foreach (var oldStr in values)
    {
        sb.Append("(");
        sb.Append(Regex.Escape(oldStr));
        sb.Append(")|");
    }
    return sb.ToString(0, sb.Length - 1);
}