feat: Integrate CastColumnExpr into PhysicalExprAdapter#20269
feat: Integrate CastColumnExpr into PhysicalExprAdapter#20269kosiew merged 3 commits intoapache:mainfrom
Conversation
github-actions
Bot
added
physical-expr
|
cc @kosiew |
adriangb
left a comment
adriangb
left a comment
There was a problem hiding this comment.
This looks like a step in the right direction to me, but I still really think we need to urgently merge CastColumnExpr into CastExpr to avoid more code paths and drift accumulating.
| /// A [`CastColumnExpr`] preserves the ordering of its child if the cast is done | ||
| /// under the same datatype family. |
There was a problem hiding this comment.
Do we have a high level SLT test for this? I'm afraid that as we refactor to e.g. fold CastColumnExpr into CastExpr we'll loose this property (which seems very valuable).
| } else if let Some(cast_expr) = | ||
| r_expr.as_any().downcast_ref::<CastColumnExpr>() |
There was a problem hiding this comment.
I think this is the unfortunate part of having two cast expressions: we have to handle both in multiple places.
@paleolimbot this makes me think that even customizing casts with a cast registry that chooses a different cast expression will be problematic. It almost feels like we either need to (1) have a single Cast expression that internally dispatches (so code snippets like this can match on a single expression) or (2) have some way to determine "is this a cast expression?". To complicate matters further: do all cast expressions share the same properties e.g. preserving order (see)
## Which issue does this PR close? <!-- We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. For example `Closes apache#123` indicates that this PR will close issue apache#123. --> - Closes apache#20163 ## Rationale for this change Ensure physical expression schema adaptation uses struct-aware casting so schema evolution for nested structs preserves field ordering, fills missing fields with NULLs, and propagates logical field metadata/nullability correctly. <!-- Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed. Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes. --> ## What changes are included in this PR? Wire schema adaptation to emit struct-aware cast expressions for column/field mismatches (including metadata and nullability), extend expression property propagation so optimizer logic continues to work, and add unit + integration coverage for struct projection and filtering under schema evolution. <!-- There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR. --> ## Are these changes tested? Yes, added new unit tests aswell <!-- We typically require tests for all PRs in order to: 1. Prevent the code from being accidentally broken by subsequent changes 2. Serve as another way to document the expected behavior of the code If tests are not included in your PR, please explain why (for example, are they covered by existing tests)? --> ## Are there any user-facing changes? No <!-- If there are user-facing changes then we may require documentation to be updated before approving the PR. --> <!-- If there are any breaking changes to public APIs, please add the `api change` label. -->
3 participants
Which issue does this PR close?
Rationale for this change
Ensure physical expression schema adaptation uses struct-aware casting so schema evolution for nested structs preserves field ordering, fills missing fields with NULLs, and propagates logical field metadata/nullability correctly.
What changes are included in this PR?
Wire schema adaptation to emit struct-aware cast expressions for column/field mismatches (including metadata and nullability), extend expression property propagation so optimizer logic continues to work, and add unit + integration coverage for struct projection and filtering under schema evolution.
Are these changes tested?
Yes, added new unit tests aswell
Are there any user-facing changes?
No