Can you please tell how to perform simple full-text search in RavenDb. The database is stored document: Movie {Name = \"Pirates of the Carribean\"}. I
Boris, Rob's answer has the right index, but it is a bit awkward for querying. You can do that using:
session.Query<Movie, Movie_ByName>()
.Search(x=>x.Name, searchTerms)
.ToList()
That will
What you are worrying about isn't anything to do with full text - by default Lucene works on an OR basis and what you want is an AND
If I were you I'd do
String[] terms = searchTerm.Split(" "); // Or whatever the string.split method is
and
.Where("Name:(" + String.Join(" AND ", terms) + ")");
Your index should look something like
public class Movie_ByName : AbstractIndexCreationTask
{
public override IndexDefinition CreateIndexDefinition()
{
return new IndexDefinitionBuilder<Movie>
{
Map = movies => from movie in movies
select new { movie.Name, market.Id },
Indexes =
{
{x => x.Name, FieldIndexing.Analyzed}
}
}
.ToIndexDefinition(DocumentStore.Conventions);
}
You don't need storage, you're not requesting the data from lucene directly at any time. You might not even want an index (You might actually want FieldIndexing.Analyzed, and might get away with just using dynamic queries here)
Up to you though.
Here's how I acheived an "ANDing" term search.
First, make sure that your field is indexed and analyzed:
public class MyIndex: AbstractIndexCreationTask<MyDocument>
{
public MyIndex()
{
Map = docs => from d in docs
select new { d.MyTextField };
Index(x => x.MyTextField, FieldIndexing.Analyzed);
}
}
Then query from the client:
var query = session.Query<MyDocument, MyIndex>();
query = theSearchText
.Split(new[] {" "}, StringSplitOptions.RemoveEmptyEntries)
.Aggregate(query, (q, term) =>
q.Search(x => x.MyTextField, term, options: SearchOptions.And));