yihua opened a new pull request, #20180: URL: https://github.com/apache/hudi/pull/20180
### Describe the issue this Pull Request addresses part of #16803 Design proposal for a Spark write path that keeps the input as Dataset/InternalRow operations end to end, with one owned routing shuffle doubling as the workload profiler and the existing row-native handles reused below the seam. The number is claimed in #20179; the proof of concept is #20166. ### Summary and Changelog Adds `rfc/rfc-115/rfc-115.md` and links its row in `rfc/README.md`. The body covers the problem and the proposal, the Delta Lake reference and why the write must own the physical plan, the architecture, index tagging and routing, the handle seam, the two profiling strategies, the entry gate, risks, open questions, the phased rollout with the status of each phase, and the test plan. Appendices hold the operation matrix, the proof-of-concept notes, the DataFrame-native record-index join sketch and the current parity tail. ### Impact Documentation only. ### Risk Level none ### Documentation Update none ### Contributor's checklist - [ ] Read through [contributor's guide](https://hudi.apache.org/contribute/how-to-contribute) - [ ] Enough context is provided in the sections above - [ ] Adequate tests were added if applicable -- 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]
