abstractdog commented on PR #6758: URL: https://github.com/apache/hive/pull/6758#issuecomment-5567313232
benchmark results: different filter selectivity, done for plain and dictionary types: ``` ┌────────┬────────────────┬────────────────┬───────────────┬────────────────┐ │ filter │ dict × on │ dict × off │ plain × on │ plain × off │ ├────────┼────────────────┼────────────────┼───────────────┼────────────────┤ │ none │ baseline 10681 │ baseline 10627 │ baseline 9875 │ baseline 10013 │ ├────────┼────────────────┼────────────────┼───────────────┼────────────────┤ │ 10 │ −24.4% │ −20.8% │ −5.6% │ −4.2% │ ├────────┼────────────────┼────────────────┼───────────────┼────────────────┤ │ 50 │ −17.8% │ −18.2% │ −4.7% │ −3.7% │ ├────────┼────────────────┼────────────────┼───────────────┼────────────────┤ │ 90 │ +1.2% │ −7.8% │ −0.7% │ −8.5% │ └────────┴────────────────┴────────────────┴───────────────┴────────────────┘ ``` Dict path: the bulk-skip fast path pays. At 10% pass (skip 90% of rows), reject runs coalesce into single RunLengthBitPackingHybridDecoder.skipInts calls and the bench saves ~24% wall-clock. At 50% pass it's still ~18%. At 90% pass, the gap collapses because most rows still materialize. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
