The YDB team has introduced full-text search with ranking and hybrid search capabilities in the 26.3 release. Hybrid search allows users to combine full-text and vector search methods within a single SQL query. The update also includes improvements to the filtered vector index functionality.
The developers explain that full-text search is effective for exact matches, such as specific document numbers, while vector search is better suited for semantic understanding of content. By integrating both, the system can handle complex queries that require both precision and contextual relevance.
A live webinar titled 'Quickly find what you need with YDB' is scheduled for October 15. During the session, the team will demonstrate how to build a search application for documents, tickets, and dialogues, integrating it with an AI assistant via RAG. All features discussed are available in the free tier.
