What problem does it solve? Writing correct Databricks SQL requires knowing which features exist, which runtime versions support them, and which syntax applies to SQL warehouses versus clusters. This Skill provides accurate syntax, version requirements, and production-ready patterns for DBSQL so you avoid trial-and-error with unfamiliar functions. ## Core Features & Use Cases - SQL Scripting and Procedures: Write procedural SQL with BEGIN...END blocks, DECLARE variables, stored procedures with error handlers, recursive CTEs, and multi-statement transactions. - AI Functions in SQL: Use ai_query, ai_classify, ai_extract, ai_mask, ai_forecast, and vector_search to enrich data with LLMs directly from SQL, plus http_request for external API calls. - Geospatial and Collation Support: Apply 39 H3 indexing functions, 80+ ST spatial functions, and ICU collations for proximity search, spatial joins, and case-insensitive queries. - Use Case: A data engineer needs to classify 10,000 support tickets by category and sentiment. The Skill provides the exact ai_classify and ai_analyze_sentiment syntax, reminds them to use a serverless SQL warehouse, and suggests LIMIT during development to control costs. ## Quick Start Ask the agent to write a Databricks SQL query that classifies support tickets by category using ai_classify on a serverless SQL warehouse.
