POC Optimize near-unique string view group values by Rachelint · Pull Request #23384 · apache/datafusion

Rachelint · 2026-07-08T10:38:16Z

Which issue does this PR close?

N/A

Rationale for this change

Final aggregation commonly receives near-unique StringView group keys from partial aggregation. For those inputs, copying every non-inline string into compact buffers adds overhead while providing little compaction benefit.

What changes are included in this PR?

Add a GroupValues::mark_near_unique hook.
Add final-stage sampling for single-column Utf8View group keys: after sampling 100,000 rows, mark group values near-unique when group_values.len() / sampled_rows > 0.8.
Teach single-column StringView group values to retain cloned input buffers in near-unique mode and remap appended views to those buffers.
Add coverage verifying near-unique StringView interning retains the input data buffer.

Are these changes tested?

Yes:

cargo fmt --all
cargo test -p datafusion-physical-expr-common binary_view_map::tests::test_bytes_view_map_mark_near_unique_retains_input_buffer
cargo test -p datafusion-physical-plan group_values
cargo clippy --all-targets --all-features -- -D warnings

Are there any user-facing changes?

No. This is an internal performance optimization.

Rachelint · 2026-07-08T10:52:44Z

run benchmarks clickbench_partitioned

adriangbot · 2026-07-08T10:54:55Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4914039528-903-ds9h5 6.12.85+ #1 SMP Mon May 11 08:17:35 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing mark-near-unique-to-group-values (ae74b26) to f34a676 (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-07-08T11:16:49Z

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Details

Comparing HEAD and mark-near-unique-to-group-values
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃      mark-near-unique-to-group-values ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.24 / 3.93 ±5.32 / 14.57 ms │          1.19 / 3.98 ±5.47 / 14.93 ms │     no change │
│ QQuery 1  │        13.23 / 13.54 ±0.16 / 13.65 ms │        12.71 / 13.27 ±0.32 / 13.64 ms │     no change │
│ QQuery 2  │        37.02 / 37.27 ±0.25 / 37.60 ms │        36.92 / 37.21 ±0.26 / 37.63 ms │     no change │
│ QQuery 3  │        30.91 / 31.75 ±0.90 / 33.50 ms │        31.04 / 31.45 ±0.39 / 32.12 ms │     no change │
│ QQuery 4  │     223.73 / 226.41 ±2.18 / 230.40 ms │     221.49 / 225.32 ±2.52 / 228.87 ms │     no change │
│ QQuery 5  │     272.52 / 274.46 ±2.29 / 278.61 ms │     264.73 / 269.13 ±3.00 / 272.49 ms │     no change │
│ QQuery 6  │           1.31 / 1.46 ±0.22 / 1.89 ms │           1.26 / 1.41 ±0.22 / 1.84 ms │     no change │
│ QQuery 7  │        14.78 / 14.97 ±0.14 / 15.17 ms │        14.27 / 14.43 ±0.13 / 14.65 ms │     no change │
│ QQuery 8  │    320.29 / 333.09 ±15.23 / 362.79 ms │     325.60 / 331.19 ±4.91 / 337.59 ms │     no change │
│ QQuery 9  │    450.33 / 463.56 ±12.88 / 485.43 ms │     456.23 / 465.18 ±7.77 / 479.43 ms │     no change │
│ QQuery 10 │        71.21 / 73.21 ±2.38 / 77.84 ms │        70.76 / 71.61 ±0.57 / 72.12 ms │     no change │
│ QQuery 11 │        82.52 / 85.41 ±3.20 / 91.64 ms │        83.33 / 84.36 ±0.95 / 86.12 ms │     no change │
│ QQuery 12 │     267.37 / 270.98 ±3.38 / 276.54 ms │     264.80 / 271.56 ±7.12 / 281.69 ms │     no change │
│ QQuery 13 │    367.72 / 382.91 ±17.71 / 416.97 ms │    366.38 / 379.97 ±11.43 / 394.88 ms │     no change │
│ QQuery 14 │     289.68 / 298.08 ±8.70 / 314.92 ms │     282.69 / 286.80 ±3.78 / 293.28 ms │     no change │
│ QQuery 15 │     271.62 / 277.13 ±3.43 / 282.24 ms │     268.70 / 274.71 ±3.71 / 279.90 ms │     no change │
│ QQuery 16 │    610.27 / 632.44 ±17.32 / 659.12 ms │     609.30 / 617.81 ±7.85 / 630.59 ms │     no change │
│ QQuery 17 │     625.09 / 634.39 ±7.38 / 647.16 ms │    618.93 / 643.91 ±18.19 / 668.34 ms │     no change │
│ QQuery 18 │ 1277.63 / 1312.49 ±24.77 / 1351.97 ms │ 1272.42 / 1286.12 ±10.22 / 1302.78 ms │     no change │
│ QQuery 19 │        28.86 / 29.10 ±0.23 / 29.48 ms │       28.18 / 35.86 ±14.77 / 65.39 ms │  1.23x slower │
│ QQuery 20 │    515.08 / 528.85 ±13.30 / 546.74 ms │     521.89 / 532.87 ±9.40 / 544.65 ms │     no change │
│ QQuery 21 │     513.93 / 521.08 ±5.10 / 529.14 ms │     516.62 / 525.30 ±8.18 / 540.75 ms │     no change │
│ QQuery 22 │  995.88 / 1011.87 ±10.93 / 1029.14 ms │   992.61 / 1001.46 ±6.10 / 1007.91 ms │     no change │
│ QQuery 23 │ 3104.49 / 3148.22 ±43.02 / 3222.36 ms │ 3118.19 / 3162.10 ±38.98 / 3235.18 ms │     no change │
│ QQuery 24 │        42.39 / 47.22 ±7.88 / 62.94 ms │        41.70 / 42.73 ±1.45 / 45.59 ms │ +1.11x faster │
│ QQuery 25 │    114.50 / 124.50 ±16.17 / 156.77 ms │     114.24 / 119.05 ±5.28 / 129.24 ms │     no change │
│ QQuery 26 │        42.75 / 44.31 ±1.60 / 46.89 ms │        41.91 / 43.90 ±2.25 / 48.19 ms │     no change │
│ QQuery 27 │    672.63 / 682.90 ±10.24 / 700.43 ms │     674.23 / 681.27 ±7.71 / 695.84 ms │     no change │
│ QQuery 28 │  3070.99 / 3077.56 ±4.35 / 3084.22 ms │  3059.00 / 3066.84 ±8.36 / 3078.89 ms │     no change │
│ QQuery 29 │        41.51 / 42.33 ±0.65 / 43.39 ms │        41.54 / 46.11 ±6.12 / 57.29 ms │  1.09x slower │
│ QQuery 30 │     306.36 / 316.82 ±9.49 / 333.79 ms │     301.93 / 307.05 ±4.22 / 312.87 ms │     no change │
│ QQuery 31 │    287.65 / 303.31 ±11.80 / 320.91 ms │     291.53 / 302.54 ±6.33 / 307.64 ms │     no change │
│ QQuery 32 │   948.52 / 981.25 ±36.28 / 1049.26 ms │   949.81 / 983.66 ±20.34 / 1010.06 ms │     no change │
│ QQuery 33 │ 1479.14 / 1495.85 ±18.03 / 1530.84 ms │ 1439.51 / 1465.54 ±30.59 / 1523.36 ms │     no change │
│ QQuery 34 │ 1482.28 / 1550.55 ±38.54 / 1595.38 ms │ 1470.38 / 1497.73 ±24.27 / 1542.41 ms │     no change │
│ QQuery 35 │    280.85 / 310.78 ±45.06 / 399.36 ms │    288.97 / 309.57 ±27.54 / 361.31 ms │     no change │
│ QQuery 36 │        70.11 / 76.69 ±5.47 / 84.69 ms │        65.43 / 68.45 ±2.90 / 73.72 ms │ +1.12x faster │
│ QQuery 37 │        36.41 / 36.84 ±0.51 / 37.80 ms │        36.45 / 41.98 ±7.36 / 56.16 ms │  1.14x slower │
│ QQuery 38 │        41.39 / 45.09 ±3.26 / 51.02 ms │        42.38 / 45.99 ±4.54 / 53.93 ms │     no change │
│ QQuery 39 │     140.54 / 153.05 ±9.85 / 165.67 ms │     146.10 / 155.98 ±9.85 / 171.53 ms │     no change │
│ QQuery 40 │        14.30 / 14.62 ±0.21 / 14.89 ms │        14.75 / 15.47 ±0.81 / 16.77 ms │  1.06x slower │
│ QQuery 41 │        14.04 / 15.44 ±2.39 / 20.21 ms │        14.05 / 18.26 ±4.73 / 25.13 ms │  1.18x slower │
│ QQuery 42 │        13.33 / 13.57 ±0.24 / 13.96 ms │        13.30 / 13.75 ±0.25 / 14.05 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                               ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                               │ 19939.27ms │
│ Total Time (mark-near-unique-to-group-values)   │ 19792.92ms │
│ Average Time (HEAD)                             │   463.70ms │
│ Average Time (mark-near-unique-to-group-values) │   460.30ms │
│ Queries Faster                                  │          2 │
│ Queries Slower                                  │          5 │
│ Queries with No Change                          │         36 │
│ Queries with Failure                            │          0 │
└─────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	105.0s
Peak memory	11.2 GiB
Avg memory	4.4 GiB
CPU user	1022.4s
CPU sys	71.3s
Peak spill	0 B

clickbench_partitioned — branch

Metric	Value
Wall time	100.0s
Peak memory	10.7 GiB
Avg memory	4.4 GiB
CPU user	1010.6s
CPU sys	73.1s
Peak spill	0 B

File an issue against this benchmark runner

Rachelint · 2026-07-08T11:17:56Z

run benchmarks clickbench_partitioned

adriangbot · 2026-07-08T11:19:02Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4914228498-904-7lzfs 6.12.85+ #1 SMP Mon May 11 08:17:35 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing mark-near-unique-to-group-values (ae74b26) to f34a676 (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-07-08T11:33:03Z

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Details

Comparing HEAD and mark-near-unique-to-group-values
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃      mark-near-unique-to-group-values ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.20 / 3.90 ±5.33 / 14.57 ms │          1.24 / 4.13 ±5.63 / 15.38 ms │  1.06x slower │
│ QQuery 1  │        12.63 / 13.16 ±0.40 / 13.87 ms │        12.86 / 13.22 ±0.20 / 13.46 ms │     no change │
│ QQuery 2  │        36.69 / 37.12 ±0.27 / 37.41 ms │        37.25 / 37.44 ±0.24 / 37.91 ms │     no change │
│ QQuery 3  │        31.25 / 32.09 ±1.02 / 34.06 ms │        31.49 / 31.75 ±0.23 / 32.02 ms │     no change │
│ QQuery 4  │     224.71 / 228.12 ±3.32 / 233.41 ms │     229.20 / 232.14 ±2.12 / 234.75 ms │     no change │
│ QQuery 5  │     271.33 / 277.02 ±3.29 / 280.38 ms │     266.98 / 269.76 ±2.06 / 273.22 ms │     no change │
│ QQuery 6  │           1.29 / 1.45 ±0.23 / 1.91 ms │           1.31 / 1.46 ±0.24 / 1.94 ms │     no change │
│ QQuery 7  │        14.32 / 14.51 ±0.18 / 14.84 ms │        14.13 / 14.38 ±0.15 / 14.59 ms │     no change │
│ QQuery 8  │     321.45 / 326.55 ±4.72 / 335.11 ms │     322.65 / 326.33 ±2.56 / 329.76 ms │     no change │
│ QQuery 9  │     461.37 / 470.27 ±9.42 / 486.74 ms │     460.30 / 466.10 ±5.67 / 476.90 ms │     no change │
│ QQuery 10 │        70.38 / 72.49 ±1.49 / 74.17 ms │        72.04 / 73.18 ±0.67 / 73.82 ms │     no change │
│ QQuery 11 │        82.93 / 84.06 ±0.76 / 84.98 ms │        84.16 / 87.30 ±4.94 / 97.13 ms │     no change │
│ QQuery 12 │     265.95 / 273.48 ±4.61 / 278.36 ms │     262.09 / 268.91 ±4.11 / 272.54 ms │     no change │
│ QQuery 13 │    373.72 / 387.59 ±17.69 / 421.89 ms │    362.65 / 373.19 ±12.89 / 398.07 ms │     no change │
│ QQuery 14 │     285.51 / 292.54 ±5.82 / 301.86 ms │    283.50 / 296.73 ±12.01 / 319.30 ms │     no change │
│ QQuery 15 │     273.90 / 279.13 ±3.34 / 284.17 ms │    275.59 / 289.79 ±23.54 / 336.65 ms │     no change │
│ QQuery 16 │     621.14 / 626.13 ±4.91 / 632.38 ms │     610.27 / 622.99 ±7.53 / 629.17 ms │     no change │
│ QQuery 17 │    626.15 / 638.04 ±12.59 / 661.58 ms │    620.99 / 636.48 ±10.81 / 648.47 ms │     no change │
│ QQuery 18 │ 1279.59 / 1307.82 ±15.57 / 1325.15 ms │ 1251.74 / 1281.08 ±15.74 / 1295.57 ms │     no change │
│ QQuery 19 │        28.05 / 28.31 ±0.34 / 28.99 ms │        28.21 / 32.69 ±8.12 / 48.91 ms │  1.15x slower │
│ QQuery 20 │    518.36 / 527.71 ±10.43 / 545.18 ms │     518.76 / 525.24 ±5.72 / 535.28 ms │     no change │
│ QQuery 21 │     519.85 / 525.69 ±7.92 / 540.99 ms │     521.25 / 530.48 ±8.37 / 545.90 ms │     no change │
│ QQuery 22 │ 1023.81 / 1047.56 ±17.77 / 1067.32 ms │  989.71 / 1004.47 ±14.04 / 1026.23 ms │     no change │
│ QQuery 23 │ 3088.10 / 3128.83 ±30.87 / 3162.86 ms │ 3103.89 / 3132.88 ±23.31 / 3161.66 ms │     no change │
│ QQuery 24 │        41.15 / 46.68 ±4.10 / 51.48 ms │        41.48 / 41.71 ±0.23 / 42.06 ms │ +1.12x faster │
│ QQuery 25 │     111.56 / 115.75 ±4.57 / 124.21 ms │     112.52 / 118.60 ±6.85 / 128.78 ms │     no change │
│ QQuery 26 │        42.16 / 44.02 ±3.09 / 50.15 ms │        43.04 / 43.97 ±0.78 / 45.38 ms │     no change │
│ QQuery 27 │     675.06 / 680.53 ±5.81 / 690.94 ms │     671.53 / 683.77 ±9.46 / 699.52 ms │     no change │
│ QQuery 28 │ 3018.35 / 3064.37 ±28.41 / 3097.01 ms │ 3007.34 / 3026.89 ±12.10 / 3038.59 ms │     no change │
│ QQuery 29 │        41.37 / 45.04 ±7.11 / 59.25 ms │       41.05 / 49.32 ±10.00 / 64.29 ms │  1.10x slower │
│ QQuery 30 │     299.85 / 310.31 ±7.77 / 319.25 ms │     302.92 / 309.68 ±7.24 / 320.44 ms │     no change │
│ QQuery 31 │     299.34 / 305.50 ±8.26 / 321.23 ms │     289.29 / 293.63 ±3.53 / 300.04 ms │     no change │
│ QQuery 32 │    941.66 / 958.57 ±15.11 / 982.88 ms │  969.62 / 1025.81 ±45.34 / 1106.65 ms │  1.07x slower │
│ QQuery 33 │ 1448.05 / 1491.37 ±43.41 / 1567.81 ms │ 1399.86 / 1450.11 ±35.18 / 1501.99 ms │     no change │
│ QQuery 34 │ 1466.34 / 1512.35 ±49.10 / 1581.78 ms │ 1436.38 / 1498.22 ±39.47 / 1553.20 ms │     no change │
│ QQuery 35 │    282.42 / 317.57 ±63.46 / 444.39 ms │    273.04 / 320.03 ±52.79 / 409.30 ms │     no change │
│ QQuery 36 │        65.50 / 68.15 ±2.73 / 72.79 ms │        64.62 / 78.46 ±9.32 / 92.59 ms │  1.15x slower │
│ QQuery 37 │       36.33 / 46.76 ±11.67 / 66.12 ms │        35.17 / 37.20 ±1.90 / 40.70 ms │ +1.26x faster │
│ QQuery 38 │        43.86 / 47.95 ±2.51 / 50.61 ms │        40.45 / 43.66 ±3.94 / 51.24 ms │ +1.10x faster │
│ QQuery 39 │     149.58 / 156.42 ±6.82 / 167.71 ms │    131.61 / 156.43 ±15.23 / 178.85 ms │     no change │
│ QQuery 40 │        14.98 / 15.40 ±0.29 / 15.86 ms │        13.97 / 15.71 ±3.08 / 21.86 ms │     no change │
│ QQuery 41 │        14.39 / 18.81 ±5.36 / 25.75 ms │        13.72 / 14.01 ±0.21 / 14.36 ms │ +1.34x faster │
│ QQuery 42 │        13.80 / 16.61 ±5.47 / 27.55 ms │        13.10 / 14.52 ±2.49 / 19.49 ms │ +1.14x faster │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                               ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                               │ 19885.74ms │
│ Total Time (mark-near-unique-to-group-values)   │ 19773.87ms │
│ Average Time (HEAD)                             │   462.46ms │
│ Average Time (mark-near-unique-to-group-values) │   459.86ms │
│ Queries Faster                                  │          5 │
│ Queries Slower                                  │          5 │
│ Queries with No Change                          │         33 │
│ Queries with Failure                            │          0 │
└─────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	100.0s
Peak memory	11.5 GiB
Avg memory	4.2 GiB
CPU user	1017.1s
CPU sys	73.8s
Peak spill	0 B

clickbench_partitioned — branch

Metric	Value
Wall time	100.0s
Peak memory	10.6 GiB
Avg memory	4.5 GiB
CPU user	1004.7s
CPU sys	74.1s
Peak spill	0 B

File an issue against this benchmark runner

Optimize near-unique string view group values

ae74b26

github-actions Bot added physical-expr Changes to the physical-expr crates physical-plan Changes to the physical-plan crate labels Jul 8, 2026

Rachelint changed the title ~~Optimize near-unique string view group values~~ POC Optimize near-unique string view group values Jul 8, 2026

Rachelint marked this pull request as draft July 8, 2026 10:53

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

POC Optimize near-unique string view group values#23384

POC Optimize near-unique string view group values#23384
Rachelint wants to merge 1 commit into
apache:mainfrom
Rachelint:mark-near-unique-to-group-values

Rachelint commented Jul 8, 2026

Uh oh!

Rachelint commented Jul 8, 2026

Uh oh!

adriangbot commented Jul 8, 2026

Uh oh!

adriangbot commented Jul 8, 2026

Uh oh!

Rachelint commented Jul 8, 2026

Uh oh!

adriangbot commented Jul 8, 2026

Uh oh!

adriangbot commented Jul 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

Rachelint commented Jul 8, 2026

Which issue does this PR close?

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

Uh oh!

Rachelint commented Jul 8, 2026

Uh oh!

adriangbot commented Jul 8, 2026

Uh oh!

adriangbot commented Jul 8, 2026

Uh oh!

Rachelint commented Jul 8, 2026

Uh oh!

adriangbot commented Jul 8, 2026

Uh oh!

adriangbot commented Jul 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants