debabhishek53 opened a new pull request, #4185: URL: https://github.com/apache/gobblin/pull/4185
## Summary
Before this fix, `CURRENT_DATE` in the legacy path (no
`iceberg.partition.value.datetime.format` set)
resolved to `LocalDateTime.now()`, causing the default `yyyy-MM-dd-HH`
formatter to embed the live
clock-hour (e.g. `2026-04-06-20`) instead of the expected `-00` suffix.
This silently broke scheduled
flows whose tables store daily data in `yyyy-MM-dd-00` partitions —
resulting in 0 files discovered
at any hour other than midnight.
**Fix:** split `CURRENT_DATE` resolution by path:
- **Legacy path** (`iceberg.partition.value.datetime.format` absent) →
`LocalDate.now().atStartOfDay()` — hour is always `00`, preserving pre-PR
behavior
- **Custom format path** (`iceberg.partition.value.datetime.format` set) →
`LocalDateTime.now()` — live clock-hour is embedded for truly
hourly-partitioned tables
## Testing
- Updated `testCurrentDatePlaceholder` to assert `today + "-00"` (strict
backward compat, deterministic)
- Added `testCurrentDatePlaceholderWithCustomFormat` for the live-hour
path (regex-based to avoid clock flakiness)
## Migration Guide
- **Daily-at-midnight partitions (`yyyy-MM-dd-00`):** no change needed —
legacy default is preserved
- **Truly hourly tables needing live hour:** set
`iceberg.partition.value.datetime.format=yyyy-MM-dd-HH`
--
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]
