In my last post, I talked about a search and separating quoted objects from free-standing words. As part of that search code, I wrote a method which helps build the search query, depending upon whether an AND radio button is selected or an OR radio button. As can be seen, we have a finite set of options here – AND and OR. Smells like an enum to me. So I created an enum:

private enum SearchType { AND, OR };

Then, a snippet of code from the method with uses that enum, which is passed in as a parameter:

if (nonQuotedSearchWords != null)
{
        foreach (string searchTermWord in nonQuotedSearchWords)
        {
             queryBuilder.Append(" CONTAINS(ALL, '" + searchTermWord + "') " + Enum.GetName(typeof(SearchType), typeOfSearch));
        }
}

typeOfSearch is the actual enum of type SearchType (the enum which I defined above).

I wrote a bit of code to perform a search using Sharepoint search functionality. The first thing I implemented was the keywords AND and OR – the normal boolean flags for searches. The next thing I implemented needed a bit more consideration. I wanted any words which were surrounded by quotes to be considered as 1 block of text. For example, “this is a quote” would be a search for the words this is a quote, in that order and with each word required. So the problem that I needed to solve was to separate out those blocks of text from the other words in the search.

My approach was to populate a generic dictionary with two kinds of objects – quoted blocks of text and separate words. The separate words could easily be rounded up into a string array. The quoted blocks needed to be separated out. I did that using a regular expression. Once I had identified those blocks of text, I added each one to the dictionary, with a key that had a sequential numbering scheme:

        protected Dictionary<string, object> ParseSearchTerms(string searchTerm)
        {
            string[] words = null;
            Dictionary<string, object> items = new Dictionary<string, object>(StringComparer.Ordinal);

            //  If there's no quoted text, just add a string array to the dictionary.
            if (searchTerm.IndexOf('"') < 0)
            {
                words = searchTerm.Split(new string[] { " " }, StringSplitOptions.RemoveEmptyEntries);
                items.Add("nonQuotedWords", words);
            }
            else
            {
                //  Identify quoted text and add it to the dictionary with the key "quotedText".
                //  Remove such text from the string to be left with a bunch of words (or nothing)
                //  which will be added as a simple string array, with the key "nonQuotedWords".
                Regex regex = new Regex("\"[^\"]*\\w*\"");
                MatchCollection matches = regex.Matches(searchTerm);
                int counter = 1;
                foreach (Match match in matches)
                {
                    items.Add("quotedText" + counter++, match.Value);
                    searchTerm = searchTerm.Replace(match.Value, string.Empty);
                }

                if (searchTerm.Length > 0)
                {
                    words = searchTerm.ToString().Split(new string[] { " " }, StringSplitOptions.RemoveEmptyEntries);
                    items.Add("nonQuotedWords", words);
                }
            }

            return items;
        }