Summary
generate_data_files.py exposes --memory-limit (bytes) for TPC-DS DuckDB generation. When omitted, the tool defaults to 50% of available system RAM and logs the resolved value when --verbose is set.
TPC-H generation currently rejects an explicit --memory-limit because the limit is not yet wired through the tpchgen-cli path (and TPC-H via --use-duckdb still uses the automatic 50% default internally).
Goal
Extend memory-limit control to TPC-H so users can cap memory consistently across both benchmarks.
Summary
generate_data_files.pyexposes--memory-limit(bytes) for TPC-DS DuckDB generation. When omitted, the tool defaults to 50% of available system RAM and logs the resolved value when--verboseis set.TPC-H generation currently rejects an explicit
--memory-limitbecause the limit is not yet wired through thetpchgen-clipath (and TPC-H via--use-duckdbstill uses the automatic 50% default internally).Goal
Extend memory-limit control to TPC-H so users can cap memory consistently across both benchmarks.