Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-04-01 Thread via GitHub
zhuqi-lucas commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2768790003 Make Clickbench Q29 5X faster: https://github.com/apache/datafusion/issues/15524 -- This is an automated message from the Apache Git Service. To respond to the messa

Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-03-12 Thread via GitHub
alamb commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2717430345 I looked more closely at Q23 and I think I know what is going on. Here is a writeup - https://github.com/apache/datafusion/issues/15177 -- This is an automated message from t

Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-03-08 Thread via GitHub
qazxcdswe123 commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2708702633 > Queries like `avg(distinct a)` rely on this rule, and without it, they cannot be executed anymore. > > The following query is available on the main branch, but not

Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-03-08 Thread via GitHub
alamb commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2708274051 > On my m1 mac 16g, removing it shows no differences. So i think we can maybe remove it? Sounds good! Can you make a PR and we can get ready to make the change? Remov

Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-03-08 Thread via GitHub
qazxcdswe123 commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2708191051 > On optimizer side, I am not sure if `single_distinct_to_groupby` can really improve performance in current version (it is an old rule introduced in long long ago), maybe

Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-02-10 Thread via GitHub
Rachelint commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2648640934 For optimizer side, I suspect if `single_distinct_to_groupby` can really improve performance in current version? -- This is an automated message from the Apache Git Service.

Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-02-10 Thread via GitHub
Rachelint commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2648625729 > > I am trying a poc about support block approach by only modifying codes of group values(we also need to modifying codes of GroupAccumulatortoo in [#11943](https://github.co

Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-02-10 Thread via GitHub
alamb commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2648610440 > I am trying a poc about support block approach by only modifying codes of group values(we also need to modifying codes of GroupAccumulatortoo in https://github.com/apache/datafu

Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-02-10 Thread via GitHub
Rachelint commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2648602542 A low hanging fruit #13617, i plan to finish it in this week. And maybe it is time to push #11943 forward... I am trying a poc about support `block approach` by `

Re: [I] [EPIC] ClickBench Improvements (Vanity Benchmark) [datafusion]

2025-02-10 Thread via GitHub
alamb commented on issue #14586: URL: https://github.com/apache/datafusion/issues/14586#issuecomment-2648588773 I took a brief look at [some results](https://benchmark.clickhouse.com/#eyJzeXN0ZW0iOnsiQWxsb3lEQiI6ZmFsc2UsIkFsbG95REIgKHR1bmVkKSI6ZmFsc2UsIkF0aGVuYSAocGFydGl0aW9uZWQpIjpmYWxzZSwi