jto commented on PR #30802: URL: https://github.com/apache/beam/pull/30802#issuecomment-2042134252
Sure. I tested it on a job that consumes ~1B records (~150GB). With the Dataset API, runtime is 37min. Passing `--useDataStreamForBatch`, I killed it after 1h+ as it was clearly too slow. I tried different bundle size settings with this path and got the following execution times: | Bundle size | execution time (min) | | ----------- | ------------------: | | 50,000 | 44 | | 100,000 | 37 | | 1,000,000 | 29 | -- 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]
