A Rust-based MCP server for SQL execution and data analysis against Parquet tables using DataFusion
DRetrieves the full schema information for a specific Parquet table, including column names and data types. This tool MUST be called before generating or executing any SQL queries to ensure schema alignment. Requires the 'name' of the table as an argument. Use this to understand the data structure and to prevent invalid column references.
Lists all registered Parquet tables in the current DataFusion context. Use this tool during the data discovery phase to identify available datasets before performing any queries. Returns a JSON array of table names. This is the mandatory first step for exploring the local data lake.
Executes an analytical SQL query against the registered Parquet tables using the DataFusion engine. The query MUST use standard SQL syntax. Before calling this, ensure you have verified the table names and column schemas using 'list_tables' and 'describe_table'. A default limit of 100 rows is applied unless specified. Use this for data aggregation, filtering, and analysis tasks.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 9 | <=2025-11-25 | v2 |