CombinedFieldQuery

A Query that treats multiple fields as a single stream and scores terms as if they had been indexed in a single field whose values would be the union of the values of the provided fields.

The query works as follows:

1. Given a list of fields and weights, it pretends there is a synthetic combined field where all terms have been indexed. It computes new term and collection statistics for this combined field.
2. It uses a disjunction iterator and [IndexSearcher.similarity] to score documents.

In order for a similarity to be compatible, [Similarity.computeNorm] must be additive: the norm of the combined field is the sum of norms for each individual field. The norms must also be encoded using [SmallFloat.intToByte4]. These requirements hold for all similarities that don't customize [Similarity.computeNorm], which includes [BM25Similarity] and [DFRSimilarity]. Per-field similarities are not supported.

The query also requires that either all fields or no fields have norms enabled. Having only some fields with norms enabled can result in errors.

This query assumes that all fields share the same analyzer. Scores may not make much sense if all fields don't have the same analyzer.

The scoring is based on BM25F's simple formula described in: http://www.staff.city.ac.uk/~sb317/papers/foundations_bm25_review.pdf. This query implements the same approach but allows other similarities besides [BM25Similarity].