Repository navigation
Conversation
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #24158 +/- ##
==========================================
+ Coverage 81.02% 81.05% +0.02%
==========================================
Files 1106 1106
Lines 380718 381147 +429
Branches 380718 381147 +429
==========================================
+ Hits 308489 308923 +434
+ Misses 54001 53967 -34
- Partials 18228 18257 +29 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
|
I see, Codecov was newly added |
bvolpato
left a comment
There was a problem hiding this comment.
LGTM. Expected-field conversion cleanly preserves nested nullability across Values literals and schema round trips. One non-blocking panic-hardening note inline.
|
|
||
| match type_variation_reference { | ||
| DEFAULT_CONTAINER_TYPE_VARIATION_REF => { | ||
| Ok(ScalarValue::List(Arc::new(ListArray::new( |
There was a problem hiding this comment.
Non-blocking: could we use ListArray::try_new / LargeListArray::try_new here and propagate the error? A Values plan can declare required list items but still contain NULL; conversion then builds a nullable child array, so new unwraps Arrow validation error and panics instead of returning a DataFusion error.
|
Let me close the PR
To avoid placing significant pressure on the reviewers, I will focus on high-priority issues |
Which issue does this PR close?
Rationale for this change
Nested literal types could mismatch the declared
VirtualTableschema, causing valid Values plans to failWhat changes are included in this PR?
Are these changes tested?
Yes,
Added unit, round-trip, deprecated Values, and integration tests
Are there any user-facing changes?
Substrait Values plans now preserve nested nullability and avoid schema/literal type mismatches