EDIT
Ok I worked it out. When I was adding the terms one at a time to the PhraseQuery object it was retaining the common words. In the case it was "The".
What I have done instead is used the QueryParser object to parse the query (including the quotation marks). This drops the common words and the phrase query now works like a charm.
List<string> searchList = Regex.Matches(searchTerms, @"(?<match>\w+)|\""(?<match>[\w\s]*)""")
.Cast<Match>()
.Select(m => m.Groups["match"].Value)
.ToList();
QueryParser parser = new QueryParser(LuceneFields.BODY, Analyzer);
BooleanQuery booleanQuery = new BooleanQuery();
// go through each term
foreach (string term in searchList)
{
Query query = null;
if (term.Contains(" ")) // multi word phrase
query = parser.Parse("\"" + term + "\"");
else
query = parser.Parse(term);
if (query.ToString() != "")
booleanQuery.Add(query, BooleanClause.Occur.MUST);
}
I am creating a simple search using Lucene.NET and I am having a bit of trouble getting phrase searching to work properly as I am combining it with a boolean query.
The following code is being used to search:
List<string> searchList = Regex.Matches(searchTerms, @"(?<match>\w+)|\""(?<match>[\w\s]*)""")
.Cast<Match>()
.Select(m => m.Groups["match"].Value)
.ToList();
QueryParser parser = new QueryParser(LuceneFields.BODY, Analyzer);
BooleanQuery booleanQuery = new BooleanQuery();
// go through each term
foreach (string term in searchList)
{
Query query = null;
if (term.Contains(" ")) // multi word phrase
{
query = new PhraseQuery();
foreach (string str in term.Split(new char[] { ' ' }, StringSplitOptions.RemoveEmptyEntries))
{
((PhraseQuery)query).Add(new Term(LuceneFields.BODY, str));
}
}
else
query = parser.Parse(term);
string strQuery = query.ToString();
if (query.ToString() != "")
booleanQuery.Add(query, BooleanClause.Occur.MUST);
}
I have checked the query that is being created and it looks OK to me:
+body:"The following table"
And I have also confirmed this text is actually in the Lucene index as you can see from a search result just searching for "table"
I'm at a loss really for what could be the problem.
I have used the following code to create the index:
Directory = FSDirectory.Open(new System.IO.DirectoryInfo(IndexDirectory));
Analyzer = new StandardAnalyzer(Version);
using (IndexWriter indexWriter = new IndexWriter(Directory, Analyzer, new IndexWriter.MaxFieldLength(Int32.MaxValue)))
{
Response.Write("Adding document...");
Document document = new Document();
// Store the IDDataContent
document.Add(new Field(LuceneFields.ID, id.ToString(), Field.Store.YES, Field.Index.ANALYZED));
// store the url to the file itself
document.Add(new Field(LuceneFields.HREF, FileURL, Field.Store.YES, Field.Index.ANALYZED));
//document.Add(new Field(LuceneFields.TITLE, Article.Title, Field.Store.YES, Field.Index.ANALYZED));
// store the text of the PDF
document.Add(new Field(LuceneFields.BODY, PdfContents, Field.Store.YES, Field.Index.ANALYZED));
indexWriter.AddDocument(document);
indexWriter.Optimize();
indexWriter.Commit();
}