vivek-9941 commented on PR #18361:
URL: https://github.com/apache/iceberg/pull/18361#issuecomment-6463820635
Thank you for re-running the probe and confirming all cases pass! I've added
comprehensive test coverage for delete files as requested.
## Tests Added
Added four test cases to `TestMergeAppend` using `newRowDelta()`:
### 1. `positionDeleteBeforePromotionCommitsWithRowDelta`
Tests position delete (v2) / deletion vector (v3-v4) built before promotion,
committed via `newRowDelta().addRows().addDeletes()`.
### 2. `dataFileBeforePromotionDeleteFileAfter`
Tests data file built before promotion, delete file (position delete v2 or
deletion vector v3-v4) built after. Verifies the cross-promotion scenario.
### 3. `equalityDeleteBeforePromotionAppliedToDataAfter`
Tests equality delete built before promotion, applied to data appended
after. **Includes scan planning** to catch missing `equalityFieldIds` (as you
noted, equality deletes only fail during scan planning).
### 4. `bothDataAndDeleteFilesBeforePromotion`
Tests both data and delete files built before promotion, then new data added
after. Comprehensive scenario covering the integer → long ClassCastException
root cause.
## Test Design
- **Format version aware**: Tests use format-appropriate delete types
(position deletes for v2, deletion vectors for v3-v4)
- **Metadata-only**: Following your probe pattern, these use the core API
without reading actual file contents
- **Scan planning verification**: All tests call
`table.newScan().planFiles()` to ensure:
- Commit succeeds
- Scan planning succeeds
- `equalityFieldIds` are preserved (critical for equality deletes)
## Helper Methods
Added three helper methods following the existing `promotionFile()` pattern:
- `promotionDeleteFile()` - Creates position deletes
- `promotionDeletionVector()` - Creates deletion vectors (v3-v4) with
referenced data file
- `promotionEqualityDeleteFile()` - Creates equality deletes on the "id"
field
All helpers accept the partition spec and value, enabling pre/post-promotion
testing.
---
Let me know if you'd like any adjustments to the test coverage!
--
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]
