How to Perform Hybrid Search

What is Reciprocal Rank Fusion?

Reciprocal rank fusion is a technique to combine results from different search methods into a single result set by performing the following actions:

1. Calculate the reciprocal rank of the documents in the results.

   For each ranked document in each search result, first add the rank (r) of the document with a constant number, 60, to smooth the score (rank_constant), and then divide 1 by the sum of r and rank_constant for the reciprocal rank of the document in the results. You can't set the value of rank_constant and it defaults to 60.

   reciprocal_rank = 1 / ( r + rank_constant )

   For each method of search, apply different weights (w) to give more importance to that method of search. For each document, the weighted reciprocal rank is calculated by multiplying the weight by the reciprocal rank of the document.

   weighted_reciprocal_rank = w x reciprocal_rank

2. Combine the rank-derived and weighted scores of the documents in the results.

   For each document across all search results, add the calculated reciprocal ranks for a single score for the document.

3. Sort the results by the combined score of the documents in the results.

   Sort the documents in the results based on the combined score across the results for a single, combined ranked list of documents in the results.

About the Different Hybrid Search Use Cases

You can leverage MongoDB Vector Search to perform several types of hybrid search. Specifically, MongoDB Vector Search supports the following use cases:

• Full-text and vector search in a single query: You can combine results from different search methods, such as a semantic and a full-text search. You can use the $vectorSearch for the semantic search and the $search for the full-text search results and combine the results by using the reciprocal rank fusion technique. To learn more, see the Perform Hybrid Search with MongoDB Vector Search and MongoDB Search tutorial, which demonstrates how to perform a semantic search and full-text search against the sample_mflix.embedded_movies namespace and retrieve combined ranked results by using reciprocal rank fusion.

  Alternatively, for a more granular hybrid search where the score matters in addition to the relative ordering of results, you can use the $scoreFusion pipeline stage. To learn more, see the Perform Hybrid Search with MongoDB Vector Search and MongoDB Search tutorial, which demonstrates how to perform a semantic search and full-text search against the sample_mflix.embedded_movies namespace and retrieve input pipeline results into a final scored results set.

  While $rankFusion ranks documents based on their positions (relative ranks) in input pipelines using the Reciprocal Rank Fusion algorithm, $scoreFusion ranks documents based on scores assigned by the input pipelines, using mathematical expressions for combining the results.

  In $rankFusion, rankings are influenced by pipeline weights. In $scoreFusion, weights control the contribution of each pipeline's scores to the final result.

• Multiple vector search queries in a single query: The MongoDB $rankFusion pipeline supports multiple sub-pipelines that contain vector search queries executed against the same collection and combining their results using the reciprocal rank fusion technique. The How to Combine Multiple $vectorSearch Queries tutorial demonstrates the following types of vector search:

  • Perform a comprehensive search of your dataset for semantically similar terms in the same query.

  • Search multiple fields in your dataset to determine which fields return the best results for the query.

  • Search using embeddings from different embedding models to determine the semantic interpretation differences between the different models.

Considerations

When using the $rankFusion or $scoreFusion pipeline stage for hybrid search, consider the following.

Limitations

The following limitations apply to hybrid search using $rankFusion and $scoreFusion:

• $rankFusion is only supported on MongoDB 8.0 and later (including latest version with auto upgrades).

  Note

  You must open a support case to use $rankFusion on 8.0.X versions. When you upgrade from 8.0, you might have to pause executing $rankFusion queries.

• $rankFusion and $scoreFusion sub-pipelines can contain only the following stages:

  • $search

  • $vectorSearch

  • $match

  • $sort

  • $geoNear

• $rankFusion and $scoreFusion preserve a traceable link back to the original input document for each sub-pipeline. Therefore, it doesn't support the following:

  • $project stage

  • storedSource fields

• $rankFusion and $scoreFusion sub-pipelines run serially, not in parallel.

• $rankFusion and $scoreFusion don't support pagination.

• rankFusion can be run on Views only on clusters running MongoDB 8.0 or higher. You can't run rankFusion within a view definition or on a time series collection.

Prerequisites

To try these tutorials, you must have the following:

• A cluster with MongoDB version v8.0 or later.

• The sample_mflix database loaded into your cluster.

• mongosh to try the queries on your cluster.

  Note

  You can also try these hybrid search use cases with local Atlas deployments that you create with the Atlas CLI and in your self-managed (on-premises) deployments. To learn more, see Create a Local Atlas Deployment.