Question

I have not been able to get this nailed down. I have tried several different analyzers and they all get me close, but not exactly what I want. Using SOLR is not an option at the moment.

An examples of what I would like:

    Input: 200
    Matches:  200 E Dragon Dr.
              200 W Paragon Rd.
              200 Lick Skillet Dr.

    Input: 200 E
    Matches:  200 E Dragon Dr.
              200 E Toll Rd.

    Input: 200 E D
    Matches: 200 E Dragon Dr.

If I use the simple analyzer then it will not match on the number. The whitespace analyzer gets the desired effect with just the number, but once I add the E it does not return as I expect. What would be the best analyzer or am I using the wrong queries?

Thanks,

EDIT:

I have taken the below answer and did a ton of googling and I am getting close just using the query parser and the whitespaceanalyzer. I am just letting the query parser determine the best query and it seems to work.

Answer 1

try using a keyword analyzer and a query parser to search an address field in Lucene. I am using a MultiFieldQueryParser, but you could use a regular query parser too:

public StartsWithQuery Prefix(string prefix, string[] fields, Dictionary<string,string> filterFields = null )
        {
           if(!string.IsNullOrEmpty(prefix))
           {


               var parser = new MultiFieldQueryParser(Version.LUCENE_29, fields, new KeywordAnalyzer());
               var boolQuery = new BooleanQuery();

               boolQuery.Add(parser.Parse(prefix + "*"), BooleanClause.Occur.MUST);
               if (filterFields != null)
               {
                   foreach (var field in filterFields)
                   {
                        boolQuery.Add(new TermQuery(new Term(field.Key, field.Value)), BooleanClause.Occur.MUST);

                   }
               }

           }

            return this;
        }