ReleaseBytes
gcp Google Cloud Blog ·

AlloyDB and Cloud SQL for PostgreSQL gain native BM25 ranking in preview

blogaigcppreviewengineerretailgcp-cloud-sql
feature announcement

Google Cloud has introduced native BM25 ranking in preview for AlloyDB and Cloud SQL for PostgreSQL 17+, integrating the `pg_textsearch` extension directly. This eliminates the need for separate full-text search backends, reducing data silos and operational complexity for developers. The capability enables robust hybrid search, combining semantic vector search with precise keyword matching, which is crucial for generative AI and RAG applications. It improves relevance for specific identifiers like alphanumeric IDs and product SKUs.

  • →Native BM25 Ranking for AlloyDB and Cloud SQL
  • →Unified Hybrid Search with Vector and Keyword Indexes
Features (2) ›
  • Native BM25 Ranking for AlloyDB and Cloud SQL

    AlloyDB and Cloud SQL for PostgreSQL 17+ now support native BM25 indexing in preview, leveraging the open-source `pg_textsearch` extension. This integration eliminates the need for external full-text search backends, simplifying data management and operational overhead for robust search applications.

  • Unified Hybrid Search with Vector and Keyword Indexes

    The update enables developers to combine semantic vector search with BM25 keyword search, addressing limitations in understanding specific IDs or SKUs in AI applications. AlloyDB further introduces an out-of-the-box hybrid search User-Defined Function (UDF) that merges results using the Reciprocal Rank Fusion (RRF) algorithm for seamless, highly relevant results.

Read the original announcement →

https://cloud.google.com/blog/products/databases/native-bm25-search-in-alloydb-and-cloud-sql/

Related releases