vector_search function
Public previewLast seen at this status: 2024-09-25·Sep 2024
A SQL table function that queries an AI Search index from Databricks SQL and returns the matching rows with all the index columns. It arrived on June 10, 2024, while the product was Mosaic AI Vector Search, at first behind an access request, and its Public Preview opened to all customers on September 25, 2024. On Databricks Runtime 15.3 and above you pass query_text or query_vector, an optional num_results (default 10, at most 100) and an optional query_type of ANN or HYBRID. It cannot query Direct Vector Access indexes, does not accept filters_json or columns, and does not run on classic SQL warehouses.
Status history
- Public preview
Sources (4)
- 2026-10-07Databricks docs Public preview
- 2024-12-17Release notes
- 2024-09-25Release notes Public preview
- 2024-06-10Release notes
Status and dates are extracted from public sources only. We don't see the private AE-shared roadmap. If something looks wrong, tell us.