databricks-vector-search

Create Databricks Vector Search endpoints, indexes, and hybrid queries.

11|3|Updated Jun 10, 2025
One-click install
npx skills add https://github.com/Paldom/databricks-apps-fastapi-starter --skill databricks-vector-search-paldom
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: databricks-vector-search
Source: https://github.com/Paldom/databricks-apps-fastapi-starter/tree/main/.gemini/skills/databricks-vector-search
Command: npx skills add https://github.com/Paldom/databricks-apps-fastapi-starter --skill databricks-vector-search-paldom

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Patterns and guidance for building, deploying, and operating Databricks Vector Search endpoints and indexes, enabling fast semantic search, RAG workflows, and similarity matching.

Core Features & Use Cases

  • Create endpoints (STANDARD or STORAGE_OPTIMIZED) to host vector indexes and run queries.
  • Build Delta Sync or Direct Access indexes with managed embeddings or precomputed vectors.
  • Perform advanced queries including hybrid search, filters, and embedding-based retrieval, across Delta tables and custom data sources.

Quick Start

Create a standard endpoint, an index on it, and run a sample query using the patterns provided.

Frequently Asked Questions about databricks-vector-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a Databricks vector search index for semantic search?

To create a Databricks vector search index, you first provision a standard or storage-optimized endpoint, then build a Delta Sync or Direct Access index using managed embeddings or precomputed vectors to enable semantic retrieval.

How do I build a RAG workflow using Databricks Vector Search?

Building RAG workflows with Databricks Vector Search involves creating an endpoint, building an index on Delta tables, and running embedding-based queries to retrieve context for your generation model.

What is the difference between Delta Sync and Direct Access indexes in Databricks?

Delta Sync indexes automatically synchronize with underlying Delta tables, while Direct Access indexes allow you to manage and ingest vectors from custom data sources without relying on Delta table synchronization.

Can I perform hybrid search with filters using Databricks Vector Search?

Yes, Databricks Vector Search supports performing advanced hybrid queries that combine semantic similarity matching with metadata filters to refine retrieval results across your indexed data.

When should I use a storage-optimized endpoint versus a standard endpoint for vector search?

Use storage-optimized endpoints for vector search workloads prioritizing lower costs on large datasets, while standard endpoints provide optimized performance for higher query throughput and lower latency.

Does Databricks Vector Search support precomputed embeddings?

Yes, Databricks Vector Search supports precomputed embeddings by utilizing Direct Access indexes, allowing you to ingest and query your own vectors instead of relying on managed embedding models.