Implement arithmetic overflow error handling #17554
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add SQL-compliant overflow checking for arithmetic operations by default. Previously, DataFusion allowed numeric overflow to wrap silently, which differs from SQL standard behavior and other databases. Changes: - Add `fail_on_overflow` configuration option (defaults to true) - Update BinaryExpr to use checked arithmetic when fail_on_overflow is enabled - Modify binary() function signature to accept ExecutionProps for config access - Fix all call sites to pass ExecutionProps parameter - Add helper functions for test code to provide default ExecutionProps The behavior now matches PostgreSQL, Trino, and Snowflake: - SELECT 10000000000 * 10000000000 returns overflow error by default - Can be disabled with SET datafusion.execution.fail_on_overflow = false Fixes apache#17539
github-actions
bot
added
physical-expr
|
PR body claims this has been tested but it doesn't look like it builds successfully in the first place? |
findepi
reviewed
Sep 15, 2025
| .as_ref() | ||
| .map(|cfg| cfg.execution.fail_on_overflow) | ||
| .unwrap_or(true); | ||
| Ok(Arc::new(BinaryExpr::new(lhs, op, rhs).with_fail_on_overflow(fail_on_overflow))) |
There was a problem hiding this comment.
Let's deprecate BinaryExpr::new and make sure all usages within DataFusion codebase are updated correctly.
| } | ||
|
|
||
| /// Helper function for tests that creates a binary expression with default ExecutionProps | ||
| fn binary_test( |
There was a problem hiding this comment.
binary_test is suitable for a boolean-returning binary operators, e.g. < or =
it's not suitable for others, e.g. +, /
Suggested change
| fn binary_test( | |
| fn binary_expr( |
| use datafusion_expr::Operator; | ||
| use datafusion_physical_expr_common::physical_expr::fmt_sql; | ||
|
|
||
| /// Helper function for tests that creates a binary expression with default ExecutionProps | ||
| fn binary_test( |
There was a problem hiding this comment.
Suggested change
| fn binary_test( | |
| fn binary_expr( |
| use datafusion_expr::Operator; | ||
|
|
||
| /// Helper function for tests that provides default ExecutionProps for binary function calls | ||
| fn binary_test( |
There was a problem hiding this comment.
Suggested change
| fn binary_test( | |
| fn binary_expr( |
| rhs: Arc<dyn PhysicalExpr>, | ||
| schema: &Schema, | ||
| ) -> Result<Arc<dyn PhysicalExpr>> { | ||
| let execution_props = ExecutionProps::new(); |
There was a problem hiding this comment.
make sure that default behavior configured by ExecutionProps::new() is to fail on overflow
|
@EeshanBembi, I am curious about the performance impact of enabling overflow checking by default. Could you add criterion benchmarks? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
Add SQL-compliant overflow checking for arithmetic operations by default. Previously,
DataFusion allowed numeric overflow to wrap silently, which differs from SQL standard
behavior and other databases like PostgreSQL, Trino, and Snowflake.
Changes
fail_on_overflowconfiguration option that defaults totruefor SQL-standardbehavior
BinaryExprto use checked arithmetic operations when overflow checking is enabledbinary()function signature to acceptExecutionPropsfor configuration accessExecutionPropsparameterExecutionPropsBehavior
Before:
Testing
Fixes #17539