voonhous opened a new issue, #19443: URL: https://github.com/apache/hudi/issues/19443
Follow-up from the review of #19404. The standalone incremental relations `IncrementalRelationV1` / `IncrementalRelationV2` (reachable only via the legacy, file-group-reader-disabled streaming path through `HoodieStreamSourceV1/V2.getBatch`) carry three branches that no test in the repo reaches. The cheapest vehicle for all three is `TestLegacyParquetReadPath.testCowIncrementalReadEqualsFileGroupReader`, which already constructs both relations directly (lines 317-319). - [ ] Fallback/full-table-scan block (`IncrementalRelationV1.scala:235-272`, `IncrementalRelationV2.scala:225-259`, including the parallelized file-existence probe from #10480): set `hoodie.datasource.read.incr.fallback.fulltablescan.enable=true` after aggressive cleaning and assert both relations still equal the file-group-reader result. This block also carries the fix history of HUDI-2711, HUDI-3189 and HUDI-7003. - [ ] File-path copy of `hoodie.datasource.read.incr.path.glob` (`IncrementalRelationV1.scala:196-207`, `IncrementalRelationV2.scala:181-192`): a second, independently written implementation of the option over `regularFileIdToFullPath` / `metaBootstrapFileIdToFullPath`, including a meta-bootstrap arm the MOR relations have no equivalent of. No test sets the glob on this path (#19404 covers only the file-slice copy in `MergeOnReadIncrementalRelationV1/V2`). - [ ] `hoodie.datasource.read.incr.schema.use.end.instanttime` (`IncrementalRelationV1.scala:121-129`, `IncrementalRelationV2.scala:106-114`, introduced by HUDI-1301 / #2125): set by no test anywhere, so the `useEndInstantSchema=true` branch is uncovered. -- 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]
