Satyr09 opened a new pull request, #25087:
URL: https://github.com/apache/datafusion/pull/25087
## 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 #123`
indicates that this PR will close issue #123.
-->
- Closes apache/datafusion#25043.
## Rationale for this change
<!--
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.
Please explain the problem you are trying to solve in terms of the
user-visible
behavior, rather than the implementation.
For example, "The code in `foo.rs` doesn't handle nulls" is a symptom of the
implementation. "COUNT(DISTINCT) returns wrong results when the column
contains
nulls" is the user-visible problem.
-->
The Substrait consumer can accept decimal arithmetic while silently changing
its declared result precision and scale. For example, division declared as
decimal(21,8) becomes a native DataFusion expression returning decimal(15,6).
This PR rejects conversions that cannot preserve the declared result type.
Casting the result alone would not recover fractional digits already lost
during calculation or prevent intermediate overflow.
## What changes are included in this PR?
<!--
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.
-->
- Validate declared decimal result types when converting addition,
subtraction, multiplication, division, and modulus into native binary
expressions.
- Reject incompatible declarations with an error identifying the function
and its declared and derived types.
- Validate decimal precision and scale, including checking protobuf integers
before narrowing conversions.
- Update ten existing TPC-H fixture tests to expect rejection of
incompatible declarations.
Plans with matching result types retain their existing behavior. Plans
omitting output_type remain accepted for compatibility. Implementing
Substrait-specific decimal arithmetic is outside this change’s scope.
## What is the testing strategy for this PR?
<!--
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
Briefly describe how this PR is tested, and point to the specific tests you
added. For example: 'This new feature is covered by the `sqllogictest` cases
added in `foo.slt`'.
If this PR does not add tests, explain why. For example, if the change is
already covered by existing tests, please mention it.
You should also check the `codecov` bot reply on this PR to confirm the
changed code is exercised.
-->
Added consumer tests covering the issue’s five examples, matching
declarations, malformed types, decimal-width differences, missing declarations,
and nested mismatches.
An integration test executes native decimal arithmetic through
producer/consumer round trips, checks result types and values, and verifies
acceptance or rejection after changing the declared output types.
Local validation:
- Formatting and diff checks passed.
- All 275 Substrait tests passed; six existing tests remain ignored.
- Clippy for Substrait targets with --no-deps passed.
- Broader Clippy checking failed on an existing unnecessary semicolon in
datafusion/common/src/rounding.rs.
- Extended workspace testing encountered failures in unchanged memory-limit
tests; the run was stopped, so full workspace validation is incomplete.
## Are there any user-facing changes?
<!--
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.
-->
Yes. Imported plans whose declared decimal arithmetic result types differ
from DataFusion’s native result types now fail during conversion instead of
silently producing a different type. Native SQL arithmetic is unchanged.
--
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]
