mirror of
https://github.com/hanzoai/datastore.git
synced 2026-09-26 13:55:07 +00:00
Restructure the README loading and known-problems sections, and make convert_csv.py return a non-zero exit code (with an error on stderr) when the final record is incomplete at EOF instead of silently dropping it, so a truncated or corrupted download cannot quietly produce a partial JOB dataset. Co-authored-by: Cursor <[email protected]>
Benchmarks
SQL benchmarks for ClickHouse. Each subdirectory contains a self-contained benchmark with the following structure:
| File / Directory | Description |
|---|---|
init.sql |
Schema definitions (CREATE TABLE statements) |
settings.json |
ClickHouse settings to use when running the queries (to make it SQL-standard compliant) |
queries/ |
The query files |