Duckdb Source

Introduction #

SmartChart supports DuckDB as a data source. DuckDB is an in-process analytical database designed for fast OLAP queries. It supports reading CSV, Excel, Parquet files directly, and can query across multiple databases without data movement.

Use Cases #

  • Analytical queries on CSV/Excel/Parquet files without importing into a database
  • Cross-database joins without ETL
  • Fast aggregation queries on large datasets

Usage #

Create a DuckDB connection pool, then write SQL queries in the dataset editor.

-- Query CSV files directly
SELECT * FROM read_csv_auto('data.csv') LIMIT 10;

-- Query Excel files
SELECT * FROM read_xlsx('data.xlsx') LIMIT 10;

-- Cross-database join
SELECT a.*, b.* FROM read_csv_auto('file1.csv') a
JOIN read_csv_auto('file2.csv') b ON a.id = b.id;

Key Features #

  • No server process needed, runs in-process
  • Supports standard SQL syntax
  • Efficient columnar storage
  • Direct file reading without import