mapleFU commented on PR #38784: URL: https://github.com/apache/arrow/pull/38784#issuecomment-1824689024
@pitrou Perhaps I need some help: 1. I found an extra problem here: https://github.com/apache/arrow/pull/38784#discussion_r1403570575 . This might the root cause of https://github.com/apache/arrow/issues/38577 . It reserve more data than expected 2. I use the code below to generate test data ``` >>> from pyarrow import parquet >>> parquet.read_table("yellow_tripdata_2010-01.parquet") >>> parquet.write_table(table, 'arrow-out.parquet') ``` When I disable `Prepare` in `DecodeArrowDenseNonNull` and `DecodeArrowDenseNonNull`, the program execute faster. The file has 14863778 rows, which means 15 row-groups. I thought that might be some reason on that but I cannot make sure. So I need some help and benchmark on that... -- 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]
