SQL Query Engine
A SQL query engine is the component that takes a SQL query, figures out the most efficient way to run it, and executes it against the data. Modern analytical engines can run directly on open file formats like Parquet, without first loading data into a proprietary store.
What a query engine does
The engine does several jobs: it parses the SQL, builds and optimizes a query plan, and then executes that plan against the data as efficiently as it can. A good engine uses techniques like vectorized execution, predicate pushdown, and partition pruning to minimize work.
A major shift in recent years is that query engines decoupled from storage. Engines like DuckDB can query Parquet files directly, wherever they live, which is what makes the lakehouse and open-format approaches practical. The engine and the data are no longer the same product.
The engine is where raw storage turns into answers, and its efficiency largely determines query speed.
How Arc handles SQL Query Engine
Arc uses DuckDB as its embedded SQL query engine, running directly on Parquet. That gives you a fast, vectorized, standard-SQL engine over open files, rather than a proprietary query layer locked to a proprietary store.
Arc is a high-performance columnar database. Open Parquet on storage you own, single Go binary, production-ready in 30 seconds.