sqd opened a new pull request, #15433: URL: https://github.com/apache/iceberg/pull/15433
When enabled, records are forwarded directly from the record generator to the writer using a forward edge instead of a hash edge. This allows Flink to chain the two operators, avoiding serialization/deserialization overhead and drastically increasing throughput in high-volume pipelines. Current topology: <img width="1064" height="468" alt="image" src="https://github.com/user-attachments/assets/7d84e5d4-10b7-490d-b086-ba986afead14"; /> Same pipeline, with the new change enabled: <img width="751" height="344" alt="image" src="https://github.com/user-attachments/assets/1ba3950c-45a3-4958-8b2d-36bb0513d7ff"; /> Serdes of Flink RowData can be very expensive: <img width="4138" height="2118" alt="image" src="https://github.com/user-attachments/assets/15186edf-e3b3-417f-bfb0-e6b2efd85848"; /> -- 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]
