Lucene: increase scoring when two terms are found

Question

I'm searching for documents that contain two terms "one two" in different fields (e.g. title, content etc.). An example in PyLucene:

query = "one two"
clauses = []
for field in fields:
    clauses.append(BooleanClause.Occur.SHOULD)

query = MultiFieldQueryParser.parse(Version.LUCENE_CURRENT, query, fields, clauses, analyzer)

I would like to get all documents that contain the term ("one" or "two") and ("one" and "two") and the documents that contain both should get a higher score. When I use a query like "one and two" or "one two"~n I get only that documents.

Is there a way to boost something like multiple matches?

Thanks.

you want to boost docs where one two present in different fields? — Mysterion, Oct 02 '14 at 09:10
Nope, I want to boost docs with both terms in the same field. — Peter Clause, Oct 02 '14 at 09:29

score 0 · Answer 1 · answered Oct 18 '14 at 11:49

If you're looking for boosting by a certain value "one two", you can use the proximity query boosting provided by SpanNearQuery

http://lucene.apache.org/core/4_6_0/core/org/apache/lucene/search/spans/SpanNearQuery.html

For instance in Java you would have something like: ...

SpanTermQuery[] spanQueryClauses = new SpanTermQuery[terms.length];
for (int i = 0; i < terms.length; i++) {
    spanQueryClauses[i] = new SpanTermQuery(terms[i]);
}
SpanNearQuery spanNearQuery = new SpanNearQuery(
              spanQueryClauses, slop, true);
spanNearQuery.setBoost(2.0f);

Lucene: increase scoring when two terms are found

1 Answers1