Brtdku

Im using the Royal Mail's sample PAF file, this data has been imported to a database and the following fields indexed via my own Lucene indexer console application:

...



var doc = new Document();



doc.Add(new Field("id", item.Id.ToString(), Field.Store.YES, Field.Index.NOT_ANALYZED));

doc.Add(new Field("postcode", item.Postcode, Field.Store.YES, Field.Index.ANALYZED));

doc.Add(new Field("buildingname", item.BuildingName, Field.Store.YES, Field.Index.ANALYZED));



...

What I want to be able to do now, is provide either a partial or full postcode or buildingname, and get matches back, as long as either searched term exists loosely in each documents postcode or buildingname fields . So if the postcode/buildingname was:

TE55 5TT Test Building

If I provided "TE55 Test" I'd expect that to come back.

My search code

var fieldsToAnalyse = new { "postcode", "buildingname" };



var finalQuery = new BooleanQuery();

var parser = new MultiFieldQueryParser(Lucene.Net.Util.Version.LUCENE_29, fieldsToAnalyse, _analyzer);



string terms = searchTerm.Split(new { " " }, StringSplitOptions.RemoveEmptyEntries);



foreach (string term in terms)

{

    var formattedTerm = term.Replace("~", "");



    var formattedTermWildcard = $"+{formattedTerm}~";



    finalQuery.Add(parser.Parse(formattedTermWildcard), Occur.MUST);



}



var searcher = new IndexSearcher(_indexDirectory, true);



var hits = searcher.Search(finalQuery,10);



foreach (var hit in hits.ScoreDocs)

{

    documents.Add(searcher.Doc(hit.Doc));

}



_analyzer.Close();

searcher.Dispose();

return documents;

Whats actually happening.

value of finalQuery is:

{+(+(postcode:test~0.5 buildingname:test~0.5)) +(+(postcode:te55~0.5
buildingname:te55~0.5))}

I'm getting back addresses that have a postcode which contains "te55" but buildingname is empty. I need it to both have a postcode that contains "te55" and the building name contain the word "test".

Sidenote

If i only provide one search term, i get:

System.IndexOutOfRangeException: 'Index was outside the bounds of the
array.'

Which is also stumping me

I would recommend to create queries programmatically (not through parsing) and also from the string version i could see that your clauses are both should (no signs around them).

As a reminder - Lucene boolean syntax is following:

+ must clause

<empty> should clause

- not clause

In your case you have

postcode:te55~0.5 buildingname:te55~0.5

which is requesting to match at least one, but not forcing both.

You need to have query like this:

+postcode:te55~0.5 +buildingname:te55~0.5

The underlying problem in MultiFieldQueryParser is that by default it's making a should clauses. You need to setDefaultOperator(AND_OPERATOR) before, so you will get desired behaviour.

Some information related from Lucene.Net 3.0.3 - https://lucenenet.apache.org/docs/3.0.3/d6/d0b/class_lucene_1_1_net_1_1_query_parsers_1_1_multi_field_query_parser.html