Repository navigation
Improve performance sort TPCH q3 with Utf8Vew ( Sort-preserving merging on a single Utf8View ) #15403
Description
Is your feature request related to a problem or challenge?
See the comments,
Sort-preserving merging on a single Utf8View column will cause sort_tpch q3 slow, we need to investigate it.
And for the q 11 case it will be 20% faster with shorter string sort.
To reproduce:
cargo run --profile release-nonlto --bin dfbench -- sort-tpch -p /Users/yongting/Code/datafusion/benchmarks/data/tpch_sf10 -q 3
Describe the solution you'd like
No response
Describe alternatives you've considered
No response
Additional context
No response
Activity
[-]Investigate why Sort-preserving merging on a single `Utf8View` column will cause sort_tpch q3 slow[/-][+]Improve performance TPCG q3 with Utf8Vew ( Sort-preserving merging on a single `Utf8View` )[/+]alamb commentedon Mar 25, 2025
Thanks, added to the list on #15096
[-]Improve performance TPCG q3 with Utf8Vew ( Sort-preserving merging on a single `Utf8View` )[/-][+]Improve performance TPCH q3 with Utf8Vew ( Sort-preserving merging on a single `Utf8View` )[/+][-]Improve performance TPCH q3 with Utf8Vew ( Sort-preserving merging on a single `Utf8View` )[/-][+]Improve performance sort TPCH q3 with Utf8Vew ( Sort-preserving merging on a single `Utf8View` )[/+]zhuqi-lucas commentedon Mar 27, 2025
Submitted a PR for review, the performance has about 40% improvement for sort-tpch q3 with single stringview column.