DuckDB Internals: Why is DuckDB Fast? (Part 1)
Summary
DuckDB Internals Part 1 explains why DuckDB is fast by detailing its in-process execution, columnar storage with zonemaps, vectorized execution, and morsel-driven parallelism, plus how the storage layer handles Parquet and CSV. It also covers how the SQL-to-logical-to-physical plan path works before execution, and hints at Part 2 for execution details.