westonpace commented on issue #15220: URL: https://github.com/apache/arrow/issues/15220#issuecomment-1373981370
Can you try a similar benchmark without involving pandas? That test will, I believe, go JSONL -> Arrow -> Pandas Dataframe -> Arrow -> Parquet. The conversion to/from pandas probably does not explain the entire time but may help if you remove it. Also, is it possible the separate the timing for the reading and writing? In other words, is the bottleneck on the read half or the write half? Finally, just to confirm, did you run each benchmark multiple times in a row? Any benchmark involving file I/O is extremely sensitive to filesystem caching effects. -- 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]
