You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@datafusion.apache.org by "Dandandan (via GitHub)" <gi...@apache.org> on 2026/03/04 07:21:53 UTC
[PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Dandandan opened a new pull request, #20687:
URL: https://github.com/apache/datafusion/pull/20687
β¦pening
When a parquet file is opened, its file-level statistics (min/max per column) are now used to update the aggregate dynamic filter bounds before any data is read. This enables earlier pruning of concurrent and subsequent files.
Key changes:
- Add DynamicFilterFileStatsHandler trait for updating filter bounds from file statistics
- Implement the trait for AggrDynFilter using inclusive operators (<=/>= instead of </>) for file-stats-derived bounds to preserve correctness
- Track bound_from_data per accumulator to switch to strict operators once the accumulator confirms the bound from actual data
- Add build_predicate() to AggrDynFilter, unifying predicate construction between file stats updates and accumulator updates
- Walk the predicate tree in the parquet opener to find and update dynamic filter nodes with file-level statistics
https://claude.ai/code/session_016tPwbdpgUiYwZSQNT8onup
## Which issue does this PR close?
<!--
We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. For example `Closes #123` indicates that this PR will close issue #123.
-->
- Closes #.
## Rationale for this change
<!--
Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes.
-->
## What changes are included in this PR?
<!--
There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR.
-->
## Are these changes tested?
<!--
We typically require tests for all PRs in order to:
1. Prevent the code from being accidentally broken by subsequent changes
2. Serve as another way to document the expected behavior of the code
If tests are not included in your PR, please explain why (for example, are they covered by existing tests)?
-->
## Are there any user-facing changes?
<!--
If there are user-facing changes then we may require documentation to be updated before approving the PR.
-->
<!--
If there are any breaking changes to public APIs, please add the `api change` label.
-->
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "alamb-ghbot (via GitHub)" <gi...@apache.org>.
alamb-ghbot commented on PR #20687:
URL: https://github.com/apache/datafusion/pull/20687#issuecomment-3996732620
π€: Benchmark completed
<details><summary>Details</summary>
<p>
```
Comparing HEAD and claude_update-parquet-dynamic-filter-YNZT0
--------------------
Benchmark clickbench_extended.json
--------------------
ββββββββββββ³ββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Query β HEAD β claude_update-parquet-dynamic-filter-YNZT0 β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 2442.03 ms β 2467.64 ms β no change β
β QQuery 1 β 995.61 ms β 986.73 ms β no change β
β QQuery 2 β 1980.12 ms β 1986.18 ms β no change β
β QQuery 3 β 1124.40 ms β 1135.28 ms β no change β
β QQuery 4 β 2408.73 ms β 2377.58 ms β no change β
β QQuery 5 β 27773.75 ms β 27123.02 ms β no change β
β QQuery 6 β 4042.40 ms β 4199.52 ms β no change β
β QQuery 7 β 2874.23 ms β 2828.65 ms β no change β
ββββββββββββ΄ββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 43641.28ms β
β Total Time (claude_update-parquet-dynamic-filter-YNZT0) β 43104.61ms β
β Average Time (HEAD) β 5455.16ms β
β Average Time (claude_update-parquet-dynamic-filter-YNZT0) β 5388.08ms β
β Queries Faster β 0 β
β Queries Slower β 0 β
β Queries with No Change β 8 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
--------------------
Benchmark clickbench_partitioned.json
--------------------
βββββββββββββ³ββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β claude_update-parquet-dynamic-filter-YNZT0 β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 2.60 ms β 2.63 ms β no change β
β QQuery 1 β 50.62 ms β 51.34 ms β no change β
β QQuery 2 β 163.16 ms β 165.84 ms β no change β
β QQuery 3 β 171.64 ms β 168.80 ms β no change β
β QQuery 4 β 1065.67 ms β 1120.18 ms β 1.05x slower β
β QQuery 5 β 1322.23 ms β 1383.54 ms β no change β
β QQuery 6 β 6.46 ms β 7.60 ms β 1.18x slower β
β QQuery 7 β 55.79 ms β 55.65 ms β no change β
β QQuery 8 β 1481.07 ms β 1515.03 ms β no change β
β QQuery 9 β 1848.86 ms β 1913.78 ms β no change β
β QQuery 10 β 359.62 ms β 365.97 ms β no change β
β QQuery 11 β 408.90 ms β 414.72 ms β no change β
β QQuery 12 β 1205.55 ms β 1262.12 ms β no change β
β QQuery 13 β 1992.35 ms β 2040.96 ms β no change β
β QQuery 14 β 1221.87 ms β 1310.66 ms β 1.07x slower β
β QQuery 15 β 1226.94 ms β 1368.50 ms β 1.12x slower β
β QQuery 16 β 2591.45 ms β 2606.20 ms β no change β
β QQuery 17 β 2563.68 ms β 2607.36 ms β no change β
β QQuery 18 β 5470.01 ms β 5169.47 ms β +1.06x faster β
β QQuery 19 β 128.88 ms β 127.31 ms β no change β
β QQuery 20 β 1992.97 ms β 1954.11 ms β no change β
β QQuery 21 β 2359.51 ms β 2235.53 ms β +1.06x faster β
β QQuery 22 β 4900.36 ms β 3947.38 ms β +1.24x faster β
β QQuery 23 β 12790.84 ms β 12292.66 ms β no change β
β QQuery 24 β 201.39 ms β 204.01 ms β no change β
β QQuery 25 β 468.29 ms β 457.63 ms β no change β
β QQuery 26 β 213.20 ms β 210.35 ms β no change β
β QQuery 27 β 2936.94 ms β 2796.01 ms β no change β
β QQuery 28 β 24505.16 ms β 24433.89 ms β no change β
β QQuery 29 β 1055.03 ms β 1033.89 ms β no change β
β QQuery 30 β 1284.83 ms β 1274.36 ms β no change β
β QQuery 31 β 1401.86 ms β 1384.76 ms β no change β
β QQuery 32 β 5264.41 ms β 4523.00 ms β +1.16x faster β
β QQuery 33 β 6295.34 ms β 5899.77 ms β +1.07x faster β
β QQuery 34 β 6340.52 ms β 6336.10 ms β no change β
β QQuery 35 β 1986.51 ms β 2031.40 ms β no change β
β QQuery 36 β 193.93 ms β 191.41 ms β no change β
β QQuery 37 β 75.28 ms β 72.20 ms β no change β
β QQuery 38 β 115.69 ms β 113.98 ms β no change β
β QQuery 39 β 343.97 ms β 347.24 ms β no change β
β QQuery 40 β 41.68 ms β 41.63 ms β no change β
β QQuery 41 β 34.55 ms β 36.08 ms β no change β
β QQuery 42 β 33.21 ms β 32.13 ms β no change β
βββββββββββββ΄ββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 98172.85ms β
β Total Time (claude_update-parquet-dynamic-filter-YNZT0) β 95507.16ms β
β Average Time (HEAD) β 2283.09ms β
β Average Time (claude_update-parquet-dynamic-filter-YNZT0) β 2221.10ms β
β Queries Faster β 5 β
β Queries Slower β 4 β
β Queries with No Change β 34 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
--------------------
Benchmark tpch_mem_sf1.json
--------------------
βββββββββββββ³ββββββββββββ³βββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β claude_update-parquet-dynamic-filter-YNZT0 β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 126.79 ms β 126.92 ms β no change β
β QQuery 2 β 32.94 ms β 32.04 ms β no change β
β QQuery 3 β 39.30 ms β 40.16 ms β no change β
β QQuery 4 β 34.59 ms β 34.92 ms β no change β
β QQuery 5 β 91.06 ms β 89.68 ms β no change β
β QQuery 6 β 24.58 ms β 24.71 ms β no change β
β QQuery 7 β 160.26 ms β 149.41 ms β +1.07x faster β
β QQuery 8 β 42.30 ms β 41.96 ms β no change β
β QQuery 9 β 105.74 ms β 111.28 ms β 1.05x slower β
β QQuery 10 β 71.88 ms β 71.12 ms β no change β
β QQuery 11 β 19.45 ms β 18.55 ms β no change β
β QQuery 12 β 68.84 ms β 68.97 ms β no change β
β QQuery 13 β 54.83 ms β 52.58 ms β no change β
β QQuery 14 β 17.62 ms β 15.60 ms β +1.13x faster β
β QQuery 15 β 35.52 ms β 33.35 ms β +1.07x faster β
β QQuery 16 β 29.86 ms β 29.69 ms β no change β
β QQuery 17 β 165.48 ms β 167.35 ms β no change β
β QQuery 18 β 302.99 ms β 295.46 ms β no change β
β QQuery 19 β 55.12 ms β 53.61 ms β no change β
β QQuery 20 β 60.79 ms β 63.74 ms β no change β
β QQuery 21 β 199.12 ms β 200.59 ms β no change β
β QQuery 22 β 24.81 ms β 24.72 ms β no change β
βββββββββββββ΄ββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 1763.91ms β
β Total Time (claude_update-parquet-dynamic-filter-YNZT0) β 1746.40ms β
β Average Time (HEAD) β 80.18ms β
β Average Time (claude_update-parquet-dynamic-filter-YNZT0) β 79.38ms β
β Queries Faster β 3 β
β Queries Slower β 1 β
β Queries with No Change β 18 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "alamb-ghbot (via GitHub)" <gi...@apache.org>.
alamb-ghbot commented on PR #20687:
URL: https://github.com/apache/datafusion/pull/20687#issuecomment-3996128676
π€: Benchmark completed
<details><summary>Details</summary>
<p>
```
Comparing HEAD and claude_update-parquet-dynamic-filter-YNZT0
--------------------
Benchmark clickbench_extended.json
--------------------
ββββββββββββ³ββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β claude_update-parquet-dynamic-filter-YNZT0 β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 2446.89 ms β 2467.16 ms β no change β
β QQuery 1 β 1015.86 ms β 920.30 ms β +1.10x faster β
β QQuery 2 β 2009.56 ms β 1907.11 ms β +1.05x faster β
β QQuery 3 β 1091.63 ms β 1123.24 ms β no change β
β QQuery 4 β 2369.14 ms β 2306.55 ms β no change β
β QQuery 5 β 28045.48 ms β 27420.40 ms β no change β
β QQuery 6 β 4081.06 ms β 4182.76 ms β no change β
β QQuery 7 β 2930.74 ms β 2693.93 ms β +1.09x faster β
ββββββββββββ΄ββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 43990.36ms β
β Total Time (claude_update-parquet-dynamic-filter-YNZT0) β 43021.45ms β
β Average Time (HEAD) β 5498.80ms β
β Average Time (claude_update-parquet-dynamic-filter-YNZT0) β 5377.68ms β
β Queries Faster β 3 β
β Queries Slower β 0 β
β Queries with No Change β 5 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
--------------------
Benchmark clickbench_partitioned.json
--------------------
βββββββββββββ³ββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β claude_update-parquet-dynamic-filter-YNZT0 β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 2.55 ms β 2.59 ms β no change β
β QQuery 1 β 51.07 ms β 51.67 ms β no change β
β QQuery 2 β 163.40 ms β 167.96 ms β no change β
β QQuery 3 β 167.48 ms β 164.41 ms β no change β
β QQuery 4 β 1070.87 ms β 1140.06 ms β 1.06x slower β
β QQuery 5 β 1397.66 ms β 1377.26 ms β no change β
β QQuery 6 β 6.69 ms β FAIL β incomparable β
β QQuery 7 β 54.61 ms β 55.51 ms β no change β
β QQuery 8 β 1530.32 ms β 1540.31 ms β no change β
β QQuery 9 β 1941.20 ms β 1849.72 ms β no change β
β QQuery 10 β 363.47 ms β 358.88 ms β no change β
β QQuery 11 β 414.87 ms β 413.00 ms β no change β
β QQuery 12 β 1278.55 ms β 1266.25 ms β no change β
β QQuery 13 β 2111.57 ms β 2053.59 ms β no change β
β QQuery 14 β 1290.44 ms β 1258.30 ms β no change β
β QQuery 15 β 1322.95 ms β 1268.57 ms β no change β
β QQuery 16 β 2804.27 ms β 2660.52 ms β +1.05x faster β
β QQuery 17 β 2655.80 ms β 2738.96 ms β no change β
β QQuery 18 β 5978.91 ms β 5298.00 ms β +1.13x faster β
β QQuery 19 β 130.72 ms β 126.94 ms β no change β
β QQuery 20 β 2032.81 ms β 1937.98 ms β no change β
β QQuery 21 β 2267.87 ms β 2242.11 ms β no change β
β QQuery 22 β 4094.14 ms β 3927.09 ms β no change β
β QQuery 23 β 24141.78 ms β 12545.91 ms β +1.92x faster β
β QQuery 24 β 202.17 ms β 211.69 ms β no change β
β QQuery 25 β 474.07 ms β 465.41 ms β no change β
β QQuery 26 β 224.10 ms β 202.44 ms β +1.11x faster β
β QQuery 27 β 2928.37 ms β 2798.14 ms β no change β
β QQuery 28 β 24497.30 ms β 24497.41 ms β no change β
β QQuery 29 β 1046.58 ms β 1051.48 ms β no change β
β QQuery 30 β 1308.86 ms β 1279.55 ms β no change β
β QQuery 31 β 1419.66 ms β 1398.21 ms β no change β
β QQuery 32 β 5155.61 ms β 5233.83 ms β no change β
β QQuery 33 β 6300.48 ms β 5945.18 ms β +1.06x faster β
β QQuery 34 β 6621.58 ms β 6615.72 ms β no change β
β QQuery 35 β 2086.93 ms β 1997.71 ms β no change β
β QQuery 36 β 192.00 ms β 188.88 ms β no change β
β QQuery 37 β 73.01 ms β 71.71 ms β no change β
β QQuery 38 β 112.71 ms β 114.93 ms β no change β
β QQuery 39 β 337.63 ms β 343.27 ms β no change β
β QQuery 40 β 39.96 ms β 39.34 ms β no change β
β QQuery 41 β 34.28 ms β 34.13 ms β no change β
β QQuery 42 β 30.15 ms β 31.32 ms β no change β
βββββββββββββ΄ββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 110352.77ms β
β Total Time (claude_update-parquet-dynamic-filter-YNZT0) β 96965.94ms β
β Average Time (HEAD) β 2627.45ms β
β Average Time (claude_update-parquet-dynamic-filter-YNZT0) β 2308.71ms β
β Queries Faster β 5 β
β Queries Slower β 1 β
β Queries with No Change β 36 β
β Queries with Failure β 1 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββ
--------------------
Benchmark tpch_mem_sf1.json
--------------------
βββββββββββββ³ββββββββββββ³βββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β claude_update-parquet-dynamic-filter-YNZT0 β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 125.14 ms β 117.66 ms β +1.06x faster β
β QQuery 2 β 32.53 ms β 33.09 ms β no change β
β QQuery 3 β 40.24 ms β 39.44 ms β no change β
β QQuery 4 β 34.74 ms β 35.27 ms β no change β
β QQuery 5 β 89.35 ms β 90.68 ms β no change β
β QQuery 6 β 24.41 ms β 24.24 ms β no change β
β QQuery 7 β 161.52 ms β 160.09 ms β no change β
β QQuery 8 β 41.10 ms β 39.90 ms β no change β
β QQuery 9 β 108.93 ms β 112.37 ms β no change β
β QQuery 10 β 73.00 ms β 70.95 ms β no change β
β QQuery 11 β 18.18 ms β 18.37 ms β no change β
β QQuery 12 β 58.40 ms β 67.20 ms β 1.15x slower β
β QQuery 13 β 53.47 ms β 53.95 ms β no change β
β QQuery 14 β 15.40 ms β 15.40 ms β no change β
β QQuery 15 β 33.40 ms β 33.24 ms β no change β
β QQuery 16 β 30.69 ms β 28.91 ms β +1.06x faster β
β QQuery 17 β 170.58 ms β 168.64 ms β no change β
β QQuery 18 β 296.29 ms β 299.48 ms β no change β
β QQuery 19 β 54.76 ms β 51.21 ms β +1.07x faster β
β QQuery 20 β 62.15 ms β 61.76 ms β no change β
β QQuery 21 β 195.93 ms β 199.71 ms β no change β
β QQuery 22 β 23.96 ms β 24.43 ms β no change β
βββββββββββββ΄ββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 1744.17ms β
β Total Time (claude_update-parquet-dynamic-filter-YNZT0) β 1746.01ms β
β Average Time (HEAD) β 79.28ms β
β Average Time (claude_update-parquet-dynamic-filter-YNZT0) β 79.36ms β
β Queries Faster β 3 β
β Queries Slower β 1 β
β Queries with No Change β 18 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "alamb-ghbot (via GitHub)" <gi...@apache.org>.
alamb-ghbot commented on PR #20687:
URL: https://github.com/apache/datafusion/pull/20687#issuecomment-3995947460
π€ `./gh_compare_branch.sh` [gh_compare_branch.sh](https://github.com/alamb/datafusion-benchmarking/blob/main/scripts/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 claude/update-parquet-dynamic-filter-YNZT0 (945a64ca6c27956669c92b394cf96e1f4b8473d6) to 2cbee474d7a3c4352a1d21308693a7e24422f722 [diff](https://github.com/apache/datafusion/compare/2cbee474d7a3c4352a1d21308693a7e24422f722..945a64ca6c27956669c92b394cf96e1f4b8473d6) using: tpch_mem clickbench_partitioned clickbench_extended
Results will be posted here when complete
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "alamb-ghbot (via GitHub)" <gi...@apache.org>.
alamb-ghbot commented on PR #20687:
URL: https://github.com/apache/datafusion/pull/20687#issuecomment-3995768774
π€ `./gh_compare_branch.sh` [gh_compare_branch.sh](https://github.com/alamb/datafusion-benchmarking/blob/main/scripts/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 claude/update-parquet-dynamic-filter-YNZT0 (0c5008d643c321b9fd8052dc44595b48d42852da) to 2cbee474d7a3c4352a1d21308693a7e24422f722 [diff](https://github.com/apache/datafusion/compare/2cbee474d7a3c4352a1d21308693a7e24422f722..0c5008d643c321b9fd8052dc44595b48d42852da) using: tpch_mem clickbench_partitioned clickbench_extended
Results will be posted here when complete
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #20687:
URL: https://github.com/apache/datafusion/pull/20687#issuecomment-3996527976
run benchmarks
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #20687:
URL: https://github.com/apache/datafusion/pull/20687#issuecomment-3995768057
run benchmarks
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan closed pull request #20687: Update aggregate dynamic filter from parquet file stats during file oβ¦
URL: https://github.com/apache/datafusion/pull/20687
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "alamb-ghbot (via GitHub)" <gi...@apache.org>.
alamb-ghbot commented on PR #20687:
URL: https://github.com/apache/datafusion/pull/20687#issuecomment-3995919554
π€: Benchmark completed
<details><summary>Details</summary>
<p>
```
Comparing HEAD and claude_update-parquet-dynamic-filter-YNZT0
--------------------
Benchmark clickbench_extended.json
--------------------
ββββββββββββ³ββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β claude_update-parquet-dynamic-filter-YNZT0 β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 2514.57 ms β 2473.21 ms β no change β
β QQuery 1 β 985.62 ms β 960.98 ms β no change β
β QQuery 2 β 2043.06 ms β 1900.71 ms β +1.07x faster β
β QQuery 3 β 1111.73 ms β 1075.57 ms β no change β
β QQuery 4 β 2412.13 ms β 2405.06 ms β no change β
β QQuery 5 β 26832.25 ms β 27105.86 ms β no change β
β QQuery 6 β 4029.28 ms β 4204.32 ms β no change β
β QQuery 7 β 2995.02 ms β 3092.93 ms β no change β
ββββββββββββ΄ββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 42923.66ms β
β Total Time (claude_update-parquet-dynamic-filter-YNZT0) β 43218.64ms β
β Average Time (HEAD) β 5365.46ms β
β Average Time (claude_update-parquet-dynamic-filter-YNZT0) β 5402.33ms β
β Queries Faster β 1 β
β Queries Slower β 0 β
β Queries with No Change β 7 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
--------------------
Benchmark clickbench_partitioned.json
--------------------
βββββββββββββ³ββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β claude_update-parquet-dynamic-filter-YNZT0 β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 2.58 ms β 2.60 ms β no change β
β QQuery 1 β 53.35 ms β 50.68 ms β +1.05x faster β
β QQuery 2 β 164.12 ms β 163.21 ms β no change β
β QQuery 3 β 172.36 ms β 165.02 ms β no change β
β QQuery 4 β 1139.55 ms β 1124.25 ms β no change β
β QQuery 5 β 1349.88 ms β 1374.60 ms β no change β
β QQuery 6 β 5.88 ms β FAIL β incomparable β
β QQuery 7 β 56.01 ms β 55.81 ms β no change β
β QQuery 8 β 1536.95 ms β 1491.06 ms β no change β
β QQuery 9 β 1843.38 ms β 1910.93 ms β no change β
β QQuery 10 β 359.87 ms β 364.84 ms β no change β
β QQuery 11 β 414.89 ms β 414.00 ms β no change β
β QQuery 12 β 1265.72 ms β 1269.00 ms β no change β
β QQuery 13 β 2028.61 ms β 2035.51 ms β no change β
β QQuery 14 β 1259.18 ms β 1308.80 ms β no change β
β QQuery 15 β 1267.97 ms β 1309.61 ms β no change β
β QQuery 16 β 2693.09 ms β 2625.51 ms β no change β
β QQuery 17 β 2656.62 ms β 2647.27 ms β no change β
β QQuery 18 β 6010.10 ms β 5017.51 ms β +1.20x faster β
β QQuery 19 β 130.32 ms β 125.49 ms β no change β
β QQuery 20 β 2026.60 ms β 1952.35 ms β no change β
β QQuery 21 β 2330.43 ms β 2286.09 ms β no change β
β QQuery 22 β 4114.62 ms β 3857.60 ms β +1.07x faster β
β QQuery 23 β 30426.93 ms β 11807.82 ms β +2.58x faster β
β QQuery 24 β 211.75 ms β 218.85 ms β no change β
β QQuery 25 β 459.72 ms β 459.96 ms β no change β
β QQuery 26 β 223.44 ms β 200.31 ms β +1.12x faster β
β QQuery 27 β 2941.31 ms β 2850.65 ms β no change β
β QQuery 28 β 24934.33 ms β 23783.73 ms β no change β
β QQuery 29 β 1011.99 ms β 1074.02 ms β 1.06x slower β
β QQuery 30 β 1233.76 ms β 1309.53 ms β 1.06x slower β
β QQuery 31 β 1331.97 ms β 1392.26 ms β no change β
β QQuery 32 β 4458.48 ms β 4399.59 ms β no change β
β QQuery 33 β 5515.25 ms β 5688.79 ms β no change β
β QQuery 34 β 5633.12 ms β 5958.14 ms β 1.06x slower β
β QQuery 35 β 1872.31 ms β 2048.84 ms β 1.09x slower β
β QQuery 36 β 190.25 ms β 193.29 ms β no change β
β QQuery 37 β 74.14 ms β 71.44 ms β no change β
β QQuery 38 β 111.63 ms β 114.31 ms β no change β
β QQuery 39 β 343.07 ms β 351.91 ms β no change β
β QQuery 40 β 39.35 ms β 40.36 ms β no change β
β QQuery 41 β 34.57 ms β 34.63 ms β no change β
β QQuery 42 β 30.62 ms β 30.94 ms β no change β
βββββββββββββ΄ββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 113954.19ms β
β Total Time (claude_update-parquet-dynamic-filter-YNZT0) β 93581.08ms β
β Average Time (HEAD) β 2713.19ms β
β Average Time (claude_update-parquet-dynamic-filter-YNZT0) β 2228.12ms β
β Queries Faster β 5 β
β Queries Slower β 4 β
β Queries with No Change β 33 β
β Queries with Failure β 1 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββ
--------------------
Benchmark tpch_mem_sf1.json
--------------------
βββββββββββββ³ββββββββββββ³βββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β claude_update-parquet-dynamic-filter-YNZT0 β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 141.44 ms β 127.96 ms β +1.11x faster β
β QQuery 2 β 34.95 ms β 33.41 ms β no change β
β QQuery 3 β 40.99 ms β 38.71 ms β +1.06x faster β
β QQuery 4 β 34.36 ms β 34.91 ms β no change β
β QQuery 5 β 89.46 ms β 90.57 ms β no change β
β QQuery 6 β 24.95 ms β 24.67 ms β no change β
β QQuery 7 β 155.00 ms β 158.91 ms β no change β
β QQuery 8 β 41.78 ms β 41.50 ms β no change β
β QQuery 9 β 115.10 ms β 108.23 ms β +1.06x faster β
β QQuery 10 β 72.01 ms β 72.25 ms β no change β
β QQuery 11 β 19.86 ms β 19.14 ms β no change β
β QQuery 12 β 67.40 ms β 69.19 ms β no change β
β QQuery 13 β 53.85 ms β 51.73 ms β no change β
β QQuery 14 β 15.75 ms β 16.37 ms β no change β
β QQuery 15 β 33.82 ms β 33.15 ms β no change β
β QQuery 16 β 29.98 ms β 29.99 ms β no change β
β QQuery 17 β 162.50 ms β 170.49 ms β no change β
β QQuery 18 β 302.51 ms β 306.45 ms β no change β
β QQuery 19 β 54.88 ms β 53.45 ms β no change β
β QQuery 20 β 61.19 ms β 62.18 ms β no change β
β QQuery 21 β 206.82 ms β 200.92 ms β no change β
β QQuery 22 β 24.31 ms β 24.34 ms β no change β
βββββββββββββ΄ββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 1782.92ms β
β Total Time (claude_update-parquet-dynamic-filter-YNZT0) β 1768.51ms β
β Average Time (HEAD) β 81.04ms β
β Average Time (claude_update-parquet-dynamic-filter-YNZT0) β 80.39ms β
β Queries Faster β 3 β
β Queries Slower β 0 β
β Queries with No Change β 19 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #20687:
URL: https://github.com/apache/datafusion/pull/20687#issuecomment-3995946873
run benchmarks
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Update aggregate dynamic filter from parquet file stats during file o⦠[datafusion]
Posted by "alamb-ghbot (via GitHub)" <gi...@apache.org>.
alamb-ghbot commented on PR #20687:
URL: https://github.com/apache/datafusion/pull/20687#issuecomment-3996528567
π€ `./gh_compare_branch.sh` [gh_compare_branch.sh](https://github.com/alamb/datafusion-benchmarking/blob/main/scripts/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 claude/update-parquet-dynamic-filter-YNZT0 (6700c7b59f8fc7fa7cf87bb1469ad6fccbc299cd) to 2cbee474d7a3c4352a1d21308693a7e24422f722 [diff](https://github.com/apache/datafusion/compare/2cbee474d7a3c4352a1d21308693a7e24422f722..6700c7b59f8fc7fa7cf87bb1469ad6fccbc299cd) using: tpch_mem clickbench_partitioned clickbench_extended
Results will be posted here when complete
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org