What problem does it solve?
Databricks Genie Spaces let users pose natural language questions and receive SQL-backed responses against Unity Catalog data. This skill facilitates creating, configuring, and querying Genie Spaces without writing SQL, enabling data teams to experiment and iterate with NL prompts.
Core Features & Use Cases
- Space Management: list_genie, create_or_update_genie, get_genie, delete_genie to provision and manage Genie Spaces.
- Conversation API: ask_genie and ask_genie_followup to run queries and continue conversations with Genie.
- Supporting Tools: get_table_details and execute_sql to inspect schemas and validate queries before execution.
- Use Cases: Build a new Genie Space for a domain, curate sample questions, test queries, and integrate with the Databricks UI for exploration.
Quick Start
Use the skill to create and query Genie Spaces. Example steps:
- Create a Genie Space with a display_name, table_identifiers, description, and sample_questions. Provide sample data like customers and orders.
- Ask questions using ask_genie(space_id, question) or ask_genie_followup(space_id, conversation_id, question).