Files
mtusi2026lab2/bench/track2/index_sizes.py
T

43 lines
1.7 KiB
Python

#!/usr/bin/env python3
"""Index sizes of every variant on the SAME 121 120-row dataset (no writes in between).
python bench/track2/index_sizes.py # ~2 min, writes evidence/track2/index-sizes.csv
"""
from __future__ import annotations
import csv
import sys
from pathlib import Path
sys.path.insert(0, str(Path(__file__).parent))
import plans # noqa: E402 (same image, limits, schema and deterministic seed)
def main() -> int:
plans.start_container("2", "2g")
try:
plans.psql("".join(p.read_text(encoding="utf-8") for p in sorted((plans.ROOT / "migrations").glob("*.sql"))))
plans.psql(plans.SEED)
rows = int(plans.psql("SELECT count(*) FROM messages;"))
out = []
for variant in plans.VARIANTS:
info = plans.use_variant(variant)
secondary = sum(int(b) for name, b in info["index_bytes"].items() if name != "messages_pkey")
out.append({"variant": variant, "rows": rows, "table_bytes": info["table_bytes"],
"pkey_bytes": int(info["index_bytes"]["messages_pkey"]), "secondary_index_bytes": secondary,
"secondary_bytes_per_row": round(secondary / rows, 1), "index_build_seconds": info["index_build_s"]})
print(out[-1], flush=True)
finally:
plans.sh(["docker", "rm", "-f", "-v", plans.CLIENT, plans.NAME], check=False)
path = plans.ROOT / "evidence/track2/index-sizes.csv"
with path.open("w", newline="", encoding="utf-8") as f:
writer = csv.DictWriter(f, fieldnames=list(out[0]))
writer.writeheader()
writer.writerows(out)
print("done:", path)
return 0
if __name__ == "__main__":
sys.exit(main())