Enable parquet filter pushdown by default #19477

alamb · 2025-12-24T10:31:37Z

Which issue does this PR close?

related to TEST: enable pushdown_filters and reorder_filters by default #18873
related to Enable parquet filter pushdown (filter_pushdown) by default #3463

Rationale for this change

I am testing where we currently are with performance

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

alamb · 2025-12-24T10:31:46Z

run benchmarks

alamb-ghbot · 2025-12-24T10:31:56Z

🤖 ./gh_compare_branch.sh gh_compare_branch.sh Running
Linux aal-dev 6.14.0-1018-gcp #19~24.04.1-Ubuntu SMP Wed Sep 24 23:23:09 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
Comparing alamb/enable_pushdown (36eae88) to 67b526a diff using: tpch_mem clickbench_partitioned clickbench_extended
Results will be posted here when complete

alamb-ghbot · 2025-12-24T11:11:07Z

🤖: Benchmark completed

Details

Comparing HEAD and alamb_enable_pushdown
--------------------
Benchmark clickbench_extended.json
--------------------
┏━━━━━━━━━━━━━━┳━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━┓
┃ Query        ┃        HEAD ┃ alamb_enable_pushdown ┃         Change ┃
┡━━━━━━━━━━━━━━╇━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━┩
│ QQuery 0     │  2685.93 ms │            2651.31 ms │      no change │
│ QQuery 1     │  1073.38 ms │            1053.67 ms │      no change │
│ QQuery 2     │  2088.87 ms │            2076.62 ms │      no change │
│ QQuery 3     │  1135.31 ms │            1169.33 ms │      no change │
│ QQuery 4     │  2311.12 ms │            2333.31 ms │      no change │
│ QQuery 5     │ 28540.84 ms │           29180.44 ms │      no change │
│ QQuery 6     │  4022.77 ms │             231.17 ms │ +17.40x faster │
│ QQuery 7     │  3930.66 ms │            4171.90 ms │   1.06x slower │
└──────────────┴─────────────┴───────────────────────┴────────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                    ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                    │ 45788.88ms │
│ Total Time (alamb_enable_pushdown)   │ 42867.76ms │
│ Average Time (HEAD)                  │  5723.61ms │
│ Average Time (alamb_enable_pushdown) │  5358.47ms │
│ Queries Faster                       │          1 │
│ Queries Slower                       │          1 │
│ Queries with No Change               │          6 │
│ Queries with Failure                 │          0 │
└──────────────────────────────────────┴────────────┘
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━━━━┳━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━┓
┃ Query        ┃        HEAD ┃ alamb_enable_pushdown ┃         Change ┃
┡━━━━━━━━━━━━━━╇━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━┩
│ QQuery 0     │     2.37 ms │               2.13 ms │  +1.12x faster │
│ QQuery 1     │    50.16 ms │              52.79 ms │   1.05x slower │
│ QQuery 2     │   134.81 ms │             131.30 ms │      no change │
│ QQuery 3     │   154.45 ms │             154.20 ms │      no change │
│ QQuery 4     │  1137.52 ms │            1117.24 ms │      no change │
│ QQuery 5     │  1511.93 ms │            1476.79 ms │      no change │
│ QQuery 6     │     2.04 ms │               2.02 ms │      no change │
│ QQuery 7     │    59.58 ms │              68.59 ms │   1.15x slower │
│ QQuery 8     │  1474.01 ms │            1456.40 ms │      no change │
│ QQuery 9     │  1854.79 ms │            1834.23 ms │      no change │
│ QQuery 10    │   353.18 ms │             492.42 ms │   1.39x slower │
│ QQuery 11    │   402.12 ms │             525.34 ms │   1.31x slower │
│ QQuery 12    │  1356.90 ms │            1555.24 ms │   1.15x slower │
│ QQuery 13    │  2021.01 ms │            2244.56 ms │   1.11x slower │
│ QQuery 14    │  1244.62 ms │            1454.12 ms │   1.17x slower │
│ QQuery 15    │  1274.32 ms │            1277.69 ms │      no change │
│ QQuery 16    │  2636.20 ms │            2571.86 ms │      no change │
│ QQuery 17    │  2601.62 ms │            2559.34 ms │      no change │
│ QQuery 18    │  5918.57 ms │            4875.16 ms │  +1.21x faster │
│ QQuery 19    │   130.29 ms │             141.55 ms │   1.09x slower │
│ QQuery 20    │  1929.33 ms │            1857.46 ms │      no change │
│ QQuery 21    │  2200.83 ms │            2330.91 ms │   1.06x slower │
│ QQuery 22    │  3857.11 ms │            3729.31 ms │      no change │
│ QQuery 23    │ 20156.62 ms │            1283.72 ms │ +15.70x faster │
│ QQuery 24    │   221.23 ms │             283.74 ms │   1.28x slower │
│ QQuery 25    │   474.00 ms │             628.50 ms │   1.33x slower │
│ QQuery 26    │   216.46 ms │             340.81 ms │   1.57x slower │
│ QQuery 27    │  2762.90 ms │            2961.52 ms │   1.07x slower │
│ QQuery 28    │ 23671.71 ms │           23963.73 ms │      no change │
│ QQuery 29    │   970.01 ms │             952.03 ms │      no change │
│ QQuery 30    │  1327.99 ms │            1342.13 ms │      no change │
│ QQuery 31    │  1345.31 ms │            1348.15 ms │      no change │
│ QQuery 32    │  5652.98 ms │            4588.56 ms │  +1.23x faster │
│ QQuery 33    │  6275.28 ms │            5666.71 ms │  +1.11x faster │
│ QQuery 34    │  6526.66 ms │            5968.31 ms │  +1.09x faster │
│ QQuery 35    │  1983.97 ms │            1908.53 ms │      no change │
│ QQuery 36    │    66.42 ms │              15.94 ms │  +4.17x faster │
│ QQuery 37    │    45.91 ms │              16.16 ms │  +2.84x faster │
│ QQuery 38    │    65.21 ms │              15.26 ms │  +4.27x faster │
│ QQuery 39    │   105.50 ms │              14.18 ms │  +7.44x faster │
│ QQuery 40    │    27.96 ms │              16.61 ms │  +1.68x faster │
│ QQuery 41    │    22.42 ms │              14.93 ms │  +1.50x faster │
│ QQuery 42    │    19.97 ms │              14.33 ms │  +1.39x faster │
└──────────────┴─────────────┴───────────────────────┴────────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━┓
┃ Benchmark Summary                    ┃             ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━┩
│ Total Time (HEAD)                    │ 104246.25ms │
│ Total Time (alamb_enable_pushdown)   │  83254.49ms │
│ Average Time (HEAD)                  │   2424.33ms │
│ Average Time (alamb_enable_pushdown) │   1936.15ms │
│ Queries Faster                       │          13 │
│ Queries Slower                       │          13 │
│ Queries with No Change               │          17 │
│ Queries with Failure                 │           0 │
└──────────────────────────────────────┴─────────────┘
--------------------
Benchmark tpch_mem_sf1.json
--------------------
┏━━━━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query        ┃      HEAD ┃ alamb_enable_pushdown ┃        Change ┃
┡━━━━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1     │ 115.22 ms │             114.68 ms │     no change │
│ QQuery 2     │  29.72 ms │              29.81 ms │     no change │
│ QQuery 3     │  39.93 ms │              38.86 ms │     no change │
│ QQuery 4     │  29.27 ms │              29.63 ms │     no change │
│ QQuery 5     │  90.54 ms │              89.69 ms │     no change │
│ QQuery 6     │  19.67 ms │              19.94 ms │     no change │
│ QQuery 7     │ 252.26 ms │             242.01 ms │     no change │
│ QQuery 8     │  38.03 ms │              39.72 ms │     no change │
│ QQuery 9     │ 105.70 ms │             103.01 ms │     no change │
│ QQuery 10    │  63.88 ms │              61.48 ms │     no change │
│ QQuery 11    │  19.33 ms │              18.03 ms │ +1.07x faster │
│ QQuery 12    │  51.68 ms │              53.21 ms │     no change │
│ QQuery 13    │  47.99 ms │              51.36 ms │  1.07x slower │
│ QQuery 14    │  14.31 ms │              14.02 ms │     no change │
│ QQuery 15    │  24.23 ms │              24.83 ms │     no change │
│ QQuery 16    │  24.31 ms │              25.01 ms │     no change │
│ QQuery 17    │ 152.94 ms │             155.39 ms │     no change │
│ QQuery 18    │ 292.03 ms │             310.42 ms │  1.06x slower │
│ QQuery 19    │  37.18 ms │              38.00 ms │     no change │
│ QQuery 20    │  49.54 ms │              51.51 ms │     no change │
│ QQuery 21    │ 333.15 ms │             350.45 ms │  1.05x slower │
│ QQuery 22    │  17.68 ms │              17.40 ms │     no change │
└──────────────┴───────────┴───────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                    ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                    │ 1848.58ms │
│ Total Time (alamb_enable_pushdown)   │ 1878.46ms │
│ Average Time (HEAD)                  │   84.03ms │
│ Average Time (alamb_enable_pushdown) │   85.38ms │
│ Queries Faster                       │         1 │
│ Queries Slower                       │         3 │
│ Queries with No Change               │        18 │
│ Queries with Failure                 │         0 │
└──────────────────────────────────────┴───────────┘

adriangb · 2025-12-25T15:59:49Z

What a 🎄 🎁 ! I see mostly speedups, vastly outweighing the slowdowns, but there are still slowdowns... any idea what those cases have in common? Any way I can help brainstorm solutions?

alamb · 2025-12-27T12:09:57Z

What a 🎄 🎁 ! I see mostly speedups, vastly outweighing the slowdowns, but there are still slowdowns... any idea what those cases have in common? Any way I can help brainstorm solutions?

Thanks @adriangb -- I am not working full time the next few days but I am hoping to work on this particular item as it is something I desparately want to be done with (and I find it fun!)

I spent some time analyzing where the time is going. I haven't written up my notes yet, but I will do so (probably tomorrow as I have a long train ride).

I found a few places to improve, but nothing that will obviously get back all the time (yet). I'll kep at it

# Which issue does this PR close? - RElated to apache/datafusion#19477 # Rationale for this change While profiling apache/datafusion#19477 I noticed some additional clones we could avoid <img width="1504" height="927" alt="Screenshot 2025-12-26 at 12 03 00 PM" src="https://github.com/user-attachments/assets/958f76e2-53d0-4008-b224-d9275984cd1a" /> I doubt this will be a huge deal but it does remove some allocations int he parquet read path # What changes are included in this PR? Use `into_data` rather than `to_data` # Are these changes tested?  # Are there any user-facing changes?

alamb · 2025-12-29T12:55:05Z

I also wrote up some notes about current status and plan here:

Enable parquet filter pushdown (filter_pushdown) by default #3463 (comment)

Dandandan · 2026-01-03T12:52:54Z

run benchmark tpch

alamb-ghbot · 2026-01-03T12:53:00Z

🤖 ./gh_compare_branch.sh gh_compare_branch.sh Running
Linux aal-dev 6.14.0-1018-gcp #19~24.04.1-Ubuntu SMP Wed Sep 24 23:23:09 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
Comparing alamb/enable_pushdown (36eae88) to 67b526a diff using: tpch
Results will be posted here when complete

Dandandan · 2026-01-03T12:53:04Z

run benchmark tpcds

alamb-ghbot · 2026-01-03T13:09:04Z

🤖: Benchmark completed

Details

Comparing HEAD and alamb_enable_pushdown
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query        ┃      HEAD ┃ alamb_enable_pushdown ┃       Change ┃
┡━━━━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1     │ 190.82 ms │             199.11 ms │    no change │
│ QQuery 2     │  93.77 ms │             128.85 ms │ 1.37x slower │
│ QQuery 3     │ 125.10 ms │             192.31 ms │ 1.54x slower │
│ QQuery 4     │  75.03 ms │             124.79 ms │ 1.66x slower │
│ QQuery 5     │ 165.04 ms │             342.06 ms │ 2.07x slower │
│ QQuery 6     │  66.31 ms │             107.95 ms │ 1.63x slower │
│ QQuery 7     │ 208.69 ms │             250.14 ms │ 1.20x slower │
│ QQuery 8     │ 164.10 ms │             279.78 ms │ 1.70x slower │
│ QQuery 9     │ 222.55 ms │             389.32 ms │ 1.75x slower │
│ QQuery 10    │ 179.97 ms │             267.35 ms │ 1.49x slower │
│ QQuery 11    │  72.19 ms │             102.21 ms │ 1.42x slower │
│ QQuery 12    │ 116.14 ms │             258.92 ms │ 2.23x slower │
│ QQuery 13    │ 213.12 ms │             209.43 ms │    no change │
│ QQuery 14    │  92.02 ms │             102.09 ms │ 1.11x slower │
│ QQuery 15    │ 119.53 ms │             167.51 ms │ 1.40x slower │
│ QQuery 16    │  56.09 ms │              89.96 ms │ 1.60x slower │
│ QQuery 17    │ 263.36 ms │             278.76 ms │ 1.06x slower │
│ QQuery 18    │ 313.29 ms │             660.48 ms │ 2.11x slower │
│ QQuery 19    │ 132.03 ms │             169.52 ms │ 1.28x slower │
│ QQuery 20    │ 124.02 ms │             147.16 ms │ 1.19x slower │
│ QQuery 21    │ 253.64 ms │             308.28 ms │ 1.22x slower │
│ QQuery 22    │  43.38 ms │              65.67 ms │ 1.51x slower │
└──────────────┴───────────┴───────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                    ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                    │ 3290.19ms │
│ Total Time (alamb_enable_pushdown)   │ 4841.65ms │
│ Average Time (HEAD)                  │  149.55ms │
│ Average Time (alamb_enable_pushdown) │  220.07ms │
│ Queries Faster                       │         0 │
│ Queries Slower                       │        20 │
│ Queries with No Change               │         2 │
│ Queries with Failure                 │         0 │
└──────────────────────────────────────┴───────────┘

alamb-ghbot · 2026-01-03T13:09:07Z

🤖 ./gh_compare_branch.sh gh_compare_branch.sh Running
Linux aal-dev 6.14.0-1018-gcp #19~24.04.1-Ubuntu SMP Wed Sep 24 23:23:09 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
Comparing alamb/enable_pushdown (36eae88) to 67b526a diff using: tpcds
Results will be posted here when complete

alamb-ghbot · 2026-01-03T13:09:08Z

Benchmark script failed with exit code 1.

Last 10 lines of output:

Click to expand

BRANCH_NAME: HEAD
DATA_DIR: /home/alamb/arrow-datafusion/benchmarks/data
RESULTS_DIR: /home/alamb/arrow-datafusion/benchmarks/results/HEAD
CARGO_COMMAND: cargo run --release
PREFER_HASH_JOIN: true
***************************

Please prepare TPC-DS data first by following instructions:
  ./bench.sh data tpcds

Dandandan · 2026-01-03T14:17:00Z

--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query        ┃      HEAD ┃ alamb_enable_pushdown ┃       Change ┃
┡━━━━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1     │ 190.82 ms │             199.11 ms │    no change │
│ QQuery 2     │  93.77 ms │             128.85 ms │ 1.37x slower │
│ QQuery 3     │ 125.10 ms │             192.31 ms │ 1.54x slower │
│ QQuery 4     │  75.03 ms │             124.79 ms │ 1.66x slower │
│ QQuery 5     │ 165.04 ms │             342.06 ms │ 2.07x slower │
│ QQuery 6     │  66.31 ms │             107.95 ms │ 1.63x slower │
│ QQuery 7     │ 208.69 ms │             250.14 ms │ 1.20x slower │
│ QQuery 8     │ 164.10 ms │             279.78 ms │ 1.70x slower │
│ QQuery 9     │ 222.55 ms │             389.32 ms │ 1.75x slower │
│ QQuery 10    │ 179.97 ms │             267.35 ms │ 1.49x slower │
│ QQuery 11    │  72.19 ms │             102.21 ms │ 1.42x slower │
│ QQuery 12    │ 116.14 ms │             258.92 ms │ 2.23x slower │
│ QQuery 13    │ 213.12 ms │             209.43 ms │    no change │
│ QQuery 14    │  92.02 ms │             102.09 ms │ 1.11x slower │
│ QQuery 15    │ 119.53 ms │             167.51 ms │ 1.40x slower │
│ QQuery 16    │  56.09 ms │              89.96 ms │ 1.60x slower │
│ QQuery 17    │ 263.36 ms │             278.76 ms │ 1.06x slower │
│ QQuery 18    │ 313.29 ms │             660.48 ms │ 2.11x slower │
│ QQuery 19    │ 132.03 ms │             169.52 ms │ 1.28x slower │
│ QQuery 20    │ 124.02 ms │             147.16 ms │ 1.19x slower │
│ QQuery 21    │ 253.64 ms │             308.28 ms │ 1.22x slower │
│ QQuery 22    │  43.38 ms │              65.67 ms │ 1.51x slower │
└──────────────┴───────────┴───────────────────────┴──────────────┘

This doesn't look great....

Dandandan · 2026-01-03T14:17:21Z

run benchmark tpch

alamb-ghbot · 2026-01-03T14:52:14Z

🤖 ./gh_compare_branch.sh gh_compare_branch.sh Running
Linux aal-dev 6.14.0-1018-gcp #19~24.04.1-Ubuntu SMP Wed Sep 24 23:23:09 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
Comparing alamb/enable_pushdown (36eae88) to 67b526a diff using: tpch
Results will be posted here when complete

alamb-ghbot · 2026-01-03T15:07:11Z

🤖: Benchmark completed

Details

Comparing HEAD and alamb_enable_pushdown
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query        ┃      HEAD ┃ alamb_enable_pushdown ┃        Change ┃
┡━━━━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1     │ 192.60 ms │             197.88 ms │     no change │
│ QQuery 2     │  92.74 ms │             129.40 ms │  1.40x slower │
│ QQuery 3     │ 122.64 ms │             198.28 ms │  1.62x slower │
│ QQuery 4     │  74.81 ms │             132.29 ms │  1.77x slower │
│ QQuery 5     │ 172.26 ms │             342.66 ms │  1.99x slower │
│ QQuery 6     │  66.49 ms │             106.20 ms │  1.60x slower │
│ QQuery 7     │ 214.54 ms │             258.79 ms │  1.21x slower │
│ QQuery 8     │ 159.64 ms │             268.68 ms │  1.68x slower │
│ QQuery 9     │ 222.34 ms │             395.58 ms │  1.78x slower │
│ QQuery 10    │ 186.05 ms │             272.04 ms │  1.46x slower │
│ QQuery 11    │  74.56 ms │             100.83 ms │  1.35x slower │
│ QQuery 12    │ 116.72 ms │             250.88 ms │  2.15x slower │
│ QQuery 13    │ 222.65 ms │             203.76 ms │ +1.09x faster │
│ QQuery 14    │  92.91 ms │             100.57 ms │  1.08x slower │
│ QQuery 15    │ 118.71 ms │             170.82 ms │  1.44x slower │
│ QQuery 16    │  54.74 ms │              86.85 ms │  1.59x slower │
│ QQuery 17    │ 267.10 ms │             283.58 ms │  1.06x slower │
│ QQuery 18    │ 314.15 ms │             658.08 ms │  2.09x slower │
│ QQuery 19    │ 134.45 ms │             169.76 ms │  1.26x slower │
│ QQuery 20    │ 126.26 ms │             145.27 ms │  1.15x slower │
│ QQuery 21    │ 259.83 ms │             310.24 ms │  1.19x slower │
│ QQuery 22    │  42.69 ms │              65.44 ms │  1.53x slower │
└──────────────┴───────────┴───────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                    ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                    │ 3328.90ms │
│ Total Time (alamb_enable_pushdown)   │ 4847.86ms │
│ Average Time (HEAD)                  │  151.31ms │
│ Average Time (alamb_enable_pushdown) │  220.36ms │
│ Queries Faster                       │         1 │
│ Queries Slower                       │        20 │
│ Queries with No Change               │         1 │
│ Queries with Failure                 │         0 │
└──────────────────────────────────────┴───────────┘

@hhhizzz

# Which issue does this PR close?  - related to apache/datafusion#19477 # Rationale for this change While working on apache/datafusion#19477, and profiling ClickBench q7, I noticed that the RowSelectors was being cloned to resolve the strategy -- for a large number of selections this is expensive and shows up in the traces <img width="1724" height="1074" alt="Screenshot 2025-12-28 at 4 49 49 PM" src="https://github.com/user-attachments/assets/72c6fd22-9377-48ef-ba80-6bc03b177cf7" /> ```shell samply record -- ./datafusion-cli-alamb_enable_pushdown -f q.sql > /dev/null 2>& ``` We should change the code to avoid cloning the RowSelectors when resolving the strategy. # Changes Don't clone / allocate while resolving the strategy. I don't expect this to have a massive impact, but it did show up in the profile FYI @hhhizzz -- perhaps you could review this PR # Are these changes tested? Yes by CI # Are there any user-facing changes? small performance improvement

# Which issue does this PR close? - related to apache/datafusion#19477 # Rationale for this change I am tracking down allocations in various parts of the parquet reader (to remove them) and I ran across this in the cached reader. # What changes are included in this PR? Use `Arc::clone` to make it clear there is no deep cloning going on . I don't expect this will have any impact on actual performance # Are these changes tested? By CI # Are there any user-facing changes?

alamb · 2026-01-08T22:07:23Z

Given that the tpch queries show a clear slowdown I am going to try profiling them and see if I can find any additional issues.

I will start with a query that is relatively simple (small number of joins):

Enable pushdown by defauly

36eae88

github-actions bot added the common Related to common crate label Dec 24, 2025

adriangb mentioned this pull request Dec 25, 2025

Redesign the try_reverse_output to support more cases #19446

Merged

alamb mentioned this pull request Dec 26, 2025

Minor: avoid some clones when reading parquet apache/arrow-rs#9048

Merged

alamb mentioned this pull request Dec 29, 2025

Reduce overhead to create an Array from ArrayData (make_array) apache/arrow-rs#9061

Open

adriangb mentioned this pull request Jan 2, 2026

perf: optimize HashTableLookupExpr::evaluate #19602

Merged

alamb mentioned this pull request Jan 8, 2026

TEST: enable pushdown_filters by default #18874

Closed

Merge remote-tracking branch 'apache/main' into alamb/enable_pushdown

be8ed6f

Enable parquet filter pushdown by default #19477

Are you sure you want to change the base?

Enable parquet filter pushdown by default #19477

Conversation

alamb commented Dec 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

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!

alamb commented Dec 24, 2025

Uh oh!

alamb-ghbot commented Dec 24, 2025

Uh oh!

alamb-ghbot commented Dec 24, 2025

Uh oh!

adriangb commented Dec 25, 2025

Uh oh!

alamb commented Dec 27, 2025

Uh oh!

alamb commented Dec 29, 2025

Uh oh!

Dandandan commented Jan 3, 2026

Uh oh!

alamb-ghbot commented Jan 3, 2026

Uh oh!

Dandandan commented Jan 3, 2026

Uh oh!

alamb-ghbot commented Jan 3, 2026

Uh oh!

alamb-ghbot commented Jan 3, 2026

Uh oh!

alamb-ghbot commented Jan 3, 2026

Uh oh!

Dandandan commented Jan 3, 2026

Uh oh!

Dandandan commented Jan 3, 2026

Uh oh!

alamb-ghbot commented Jan 3, 2026

Uh oh!

alamb-ghbot commented Jan 3, 2026

Uh oh!

alamb commented Jan 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

alamb commented Dec 24, 2025 •

edited

Loading