fix: deserialization error for FilterExec (predicates with inlist)

[haohuaijin]

Which issue does this PR close?

• closes regression: inlist deserialization error #17225

Rationale for this change

fix #17225

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

[haohuaijin]

I check the code, the FliterExec predicate's column index is based on FilterExec's input schema, so when we deserialization the FilterExec, the predicate should also based on input schema.

datafusion/datafusion/physical-plan/src/filter.rs

Lines 90 to 116 in b84ddfd

	pub fn try_new(
	predicate: Arc<dyn PhysicalExpr>,
	input: Arc<dyn ExecutionPlan>,
	) -> Result<Self> {
	match predicate.data_type(input.schema().as_ref())? {
	DataType::Boolean => {
	let default_selectivity = FILTER_EXEC_DEFAULT_SELECTIVITY;
	let cache = Self::compute_properties(
	&input,
	&predicate,
	default_selectivity,
	None,
	)?;
	Ok(Self {
	predicate,
	input: Arc::clone(&input),
	metrics: ExecutionPlanMetricsSet::new(),
	default_selectivity,
	cache,
	projection: None,
	})
	}
	other => {
	plan_err!("Filter predicate must return BOOLEAN values, got {other:?}")
	}
	}
	}

and i also check the disscussion in #16744 and #16665, not sure why previous issue happen. maybe some code have problem before, but someone fix it recently

cc @alamb @XiangpengHao

[alamb]

Thank you @haohuaijin -- this is my favorite kind of bug fix (fix bug by deleting code!)

cc @XiangpengHao and @NGA-TRAN

[alamb]

Using the input schema always matches my understanding of how the FilterExec work 👍

[alamb]

I verified this test covers the code as it fails without the fix

---- cases::roundtrip_physical_plan::test_tpch_part_in_list_query_with_real_parquet_data stdout ----
Error: Internal("PhysicalExpr Column references column 'p_size' at index 1 (zero-based) but input schema only has 1 columns: [\"p_size\"]")

[alamb]

My understanding of this fix is that it could affect any predicate, not just IN lists, so I updated the PR description to match

[haohuaijin]

Thank you @alamb

[alamb]

I merged up from main to resolve a conflict

[alamb]

Thanks again @haohuaijin