Blizzara opened a new issue, #15190:
URL: https://github.com/apache/datafusion/issues/15190
### Describe the bug
It seems the Common Subexpression Elimination rule may eliminate also
WindowFunctions by moving them into a Project node, which then causes physical
planning to fail since a Project cannot actually execute those WindowFunctions.
### To Reproduce
The following test fails on current main:
```
#[tokio::test]
async fn window() -> Result<()> {
let wexpr = row_number_udwf().call(vec![]);
let plan = LogicalPlanBuilder::empty(true)
.window(vec![wexpr.clone(), wexpr.alias("aliased")])?
.build()?;
println!("plan: {}", plan.display_indent_schema());
let ctx = SessionStateBuilder::new().build();
let plan = ctx.optimize(&plan)?;
println!("optimized plan: {}", plan.display_indent_schema());
DataFrame::new(ctx, plan).collect().await?;
Ok(())
}
```
with
```
plan: WindowAggr: windowExpr=[[row_number() ROWS BETWEEN UNBOUNDED PRECEDING
AND UNBOUNDED FOLLOWING, row_number() ROWS BETWEEN UNBOUNDED PRECEDING AND
UNBOUNDED FOLLOWING AS aliased]] [row_number() ROWS BETWEEN UNBOUNDED PRECEDING
AND UNBOUNDED FOLLOWING:UInt64, aliased:UInt64]
EmptyRelation []
optimized plan: Projection: row_number() ROWS BETWEEN UNBOUNDED PRECEDING
AND UNBOUNDED FOLLOWING, aliased [row_number() ROWS BETWEEN UNBOUNDED PRECEDING
AND UNBOUNDED FOLLOWING:UInt64, aliased:UInt64]
WindowAggr: windowExpr=[[__common_expr_1 AS row_number() ROWS BETWEEN
UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING, __common_expr_1 AS aliased]]
[__common_expr_1:UInt64, row_number() ROWS BETWEEN UNBOUNDED PRECEDING AND
UNBOUNDED FOLLOWING:UInt64, aliased:UInt64]
Projection: row_number() ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED
FOLLOWING AS __common_expr_1 [__common_expr_1:UInt64]
EmptyRelation []
Error: NotImplemented("Physical plan does not support logical expression
WindowFunction(WindowFunction { fun: WindowUDF(WindowUDF { inner: RowNumber {
signature: Signature { type_signature: Nullary, volatility: Immutable } } }),
params: WindowFunctionParams { args: [], partition_by: [], order_by: [],
window_frame: WindowFrame { units: Rows, start_bound: Preceding(UInt64(NULL)),
end_bound: Following(UInt64(NULL)), is_causal: false }, null_treatment: None }
})")
```
Note how the optimize() has introduced a Project node which contains the
row_number() call.
Removing either expr from the `window()` call makes the test pass.
### Expected behavior
Test succeeds. Preferrably the Window node notices the common subexpression
and deduplicates it, but that's a bonus, as long as the plan runs.
### Additional context
_No response_
--
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: github-unsubscr...@datafusion.apache.org.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For additional commands, e-mail: github-h...@datafusion.apache.org
