danny0405 opened a new pull request, #20675: URL: https://github.com/apache/hudi/pull/20675
### Describe the issue this Pull Request addresses Issue: #20405; parent #20401. The existing SDP materialized-view scenario only covers explicitly keyed targets. Add the deferred keyless aggregate acceptance without claiming the remaining refresh/failure criteria are complete. ### Summary and Changelog - Add an unpartitioned COW aggregate MV with no primaryKey, grouping JSON input by value and computing sum(id) and count(*). - Check exact results for the aggregate and both existing keyed views after four real CLI refreshes. Cover a changed aggregate for a retained group, removed/new groups, and an empty result. - Keep explicit checks active under Python optimization and retain harness cleanup and diagnostics. - Update the runnable scenario documentation. No external code was copied. ### Impact Test and developer-documentation changes only. No runtime, API, configuration or storage-format changes. Uses current purge-style refresh semantics. ### Risk Level low. Built the current-master Spark 4.1 candidate bundle with JDK 17/Scala 2.13; reactor install and style/license checks passed. The real Spark Connect 4.1.1 CLI scenario passed under python -O: four refreshes and twelve exact-result checks. Python compilation and git diff --check passed. Metadata indexing remains disabled; this does not establish atomic refresh, timeline preservation or injected-failure coverage. ### Documentation Update Updated hudi-integ-test/spark-pipelines/MATERIALIZED_VIEWS.md with the keyless definition, input matrix and exact aggregate expectations. No website change is needed for test infrastructure. ### Contributor's checklist - [x] Read through [contributor's guide](https://hudi.apache.org/contribute/how-to-contribute) - [x] Enough context is provided in the sections above - [x] 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]
