samuelcolvin commented on issue #7845: URL: https://github.com/apache/datafusion/issues/7845#issuecomment-2370455772
I've shared the code I'm using to run these tests: https://github.com/samuelcolvin/batson-perf Unfortunately setting `datafusion.execution.parquet.pushdown_filters` to `true` makes things slightly worse: ``` code/batson-perf 130 ➤ cargo run --release Compiling jiter v0.5.0 (https://github.com/pydantic/jiter.git?branch=batson#6513df59) Compiling batson v0.5.0 (https://github.com/pydantic/jiter.git?branch=batson#6513df59) Compiling batson-perf v0.1.0 (/Users/samuel/code/batson-perf) Finished `release` profile [optimized] target(s) in 2.47s Running `target/release/batson-perf` loading data from records_1000000_zstd(1).parquet running queries... +-----------------+ | count(Int64(1)) | +-----------------+ | 105461 | +-----------------+ mode: FilterJson zstd(1), pushdown_filters: false, query took 1.123133083s +-----------------+ | count(Int64(1)) | +-----------------+ | 105461 | +-----------------+ mode: FilterBatson zstd(1), pushdown_filters: false, query took 544.314375ms code/batson-perf 0 ➤ cargo run --release Compiling batson-perf v0.1.0 (/Users/samuel/code/batson-perf) Finished `release` profile [optimized] target(s) in 1.53s Running `target/release/batson-perf` loading data from records_1000000_zstd(1).parquet running queries... +-----------------+ | count(Int64(1)) | +-----------------+ | 105461 | +-----------------+ mode: FilterJson zstd(1), pushdown_filters: true, query took 1.426721083s +-----------------+ | count(Int64(1)) | +-----------------+ | 105461 | +-----------------+ mode: FilterBatson zstd(1), pushdown_filters: true, query took 957.441ms ``` -- 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]
