Add list-aware element repartitioning - #9028
Conversation
Signed-off-by: Matt Katz <mhkatz97@gmail.com>
Signed-off-by: Matt Katz <mhkatz97@gmail.com>
Polar Signals Profiling ResultsLatest Run
Powered by Polar Signals Cloud |
Benchmarks: TPC-H SF=1 on NVME 📖Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.947x ➖, 1↑ 0↓)
datafusion / parquet (0.953x ➖, 2↑ 0↓)
duckdb / vortex-file-compressed (0.956x ➖, 1↑ 0↓)
duckdb / parquet (0.958x ➖, 2↑ 0↓)
File Size Changes (10 files changed, -43.9% overall, 0↑ 10↓)
Totals:
|
Benchmarks: PolarSignals Profiling 📖Vortex (geomean): 9.493x ❌ How to read Verdict and Engines
datafusion / vortex-file-compressed (9.493x ❌, 0↑ 10↓)
File Size Changes (1 files changed, +14.5% overall, 1↑ 0↓)
Totals:
|
Benchmarks: FineWeb NVMe 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.160x ❌, 0↑ 7↓)
datafusion / parquet (1.097x ➖, 0↑ 4↓)
duckdb / vortex-file-compressed (1.175x ❌, 0↑ 8↓)
duckdb / parquet (1.074x ➖, 0↑ 2↓)
File Size Changes (2 files changed, -46.3% overall, 0↑ 2↓)
Totals:
|
Benchmarks: TPC-DS SF=1 on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.032x ➖, 0↑ 3↓)
datafusion / parquet (1.028x ➖, 1↑ 5↓)
duckdb / vortex-file-compressed (1.032x ➖, 0↑ 5↓)
duckdb / parquet (1.020x ➖, 0↑ 2↓)
File Size Changes (25 files changed, -43.5% overall, 0↑ 25↓)
Totals:
|
🚨🚨🚨❌❌❌ SQL BENCHMARK FAILED ❌❌❌🚨🚨🚨Benchmark |
Benchmarks: TPC-H SF=10 on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.079x ➖, 0↑ 5↓)
datafusion / parquet (1.057x ➖, 0↑ 1↓)
duckdb / vortex-file-compressed (1.029x ➖, 0↑ 0↓)
duckdb / parquet (1.027x ➖, 0↑ 0↓)
File Size Changes (25 files changed, -44.0% overall, 0↑ 25↓)
Totals:
|
Benchmarks: FineWeb S3 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.141x ➖, 0↑ 2↓)
datafusion / parquet (1.036x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (1.009x ➖, 0↑ 0↓)
duckdb / parquet (1.020x ➖, 0↑ 0↓)
|
Benchmarks: Clickbench Sorted on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.943x ➖, 1↑ 0↓)
datafusion / parquet (0.955x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed (0.930x ➖, 5↑ 1↓)
duckdb / parquet (0.975x ➖, 0↑ 0↓)
File Size Changes (201 files changed, -42.8% overall, 53↑ 148↓)
Totals:
|
Benchmarks: Clickbench on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.000x ➖, 1↑ 0↓)
datafusion / parquet (1.007x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (0.988x ➖, 3↑ 0↓)
duckdb / parquet (1.005x ➖, 0↑ 0↓)
File Size Changes (101 files changed, -39.2% overall, 0↑ 101↓)
Totals:
|
Benchmarks: TPC-H SF=1 on S3 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.010x ➖, 0↑ 1↓)
datafusion / parquet (0.993x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (1.036x ➖, 0↑ 0↓)
duckdb / parquet (0.993x ➖, 0↑ 0↓)
|
Repartition list elements toward the configured byte target without splitting sublists across chunks. Integrate this into
ListLayoutStrategy.