What problem does it solve? Analysts and engineers working with Microsoft Fabric often need fast, ad-hoc access to lakehouse and warehouse data without spinning up Spark jobs or writing full pipelines. This Skill lets you query Delta tables and raw files in OneLake directly with DuckDB, either from your local machine or inside a Fabric notebook. ## Core Features & Use Cases - Local DuckDB querying: Connect to OneLake over ABFS using Azure CLI credentials and run SQL against Delta tables, CSV, Parquet, and JSON files with glob pattern support. - In-notebook attachment: Attach DuckDB to a lakehouse inside a Fabric notebook using a storage token, auto-discover all Delta tables as views, and write results back through Spark. - Common analytical patterns: Run data freshness checks, data quality validation, schema discovery, cross-table joins, and row count audits with ready-made SQL templates. - Use Case: You want to verify that yesterday's orders table loaded correctly. Use this Skill to run a freshness check and row count audit against the lakehouse in seconds, without starting a Spark session. ## Quick Start Ask the agent to use DuckDB to query the top rows of a Delta table in your Fabric lakehouse and check its data freshness.
