ewoodbury commented on PR #3255:
URL: https://github.com/apache/iceberg-rust/pull/3255#issuecomment-6091383398
Thanks @laskoviymishka! Pre-merge items are in:
1. Reader-level field-id test: `test_read_evolved_nested_fields_by_field_id`
writes a real Parquet file with an old nested schema and reads it back through
an evolved one. It fails on main with Arrow Schema Error, and here it errors
rather than falling back to position if nested field ids go missing (test was
suggested by @comphead and +1'd here).
2. Id-less fallback: kept on purpose, pinned with a test, and the comment
points at #1845.
I also applied the duplicate-id check, the plan-time guard for nested
sources read as a non-nested type, `DataInvalid` for map entries, the map
data_type asserts, and a test covering sliced-lists. #3378 is filed, with a
pointer in `get_arrow_projection_mask_with_field_ids`.
Mentioned in replies: for #3261 I'd rather close it completely in a focused
follow-up than half-close it here. That PR should cover nested defaults, the
non-null {} struct default from #2635, and the shared absent-field helper. So
this keeps the nested initial-default as `FeatureUnsupported`, and I've added a
release note for that to the description. Let me know if any final notes!
--
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]
