Count wildcard alias

[jayzhan211]

Which issue does this PR close?

• Closes Change in 46: count_all() expr_fn function now displayed as count(1) rather than count(*) #14894.

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

[jayzhan211]

    /// Return `self AS name` alias expression
    pub fn alias(self, name: impl Into<String>) -> Expr {
        Expr::Alias(Alias::new(self, None::<&str>, name.into()))
    }

Add duplicate name check here didn't help.

datafusion/datafusion/expr/src/logical_plan/builder.rs

Lines 748 to 768 in 32224b4

	let mut missing_cols: IndexSet<Column> = IndexSet::new();
	sorts.iter().try_for_each::<_, Result<()>>(|sort| {
	let columns = sort.expr.column_refs();
	missing_cols.extend(
	columns
	.into_iter()
	.filter(|c| !schema.has_column(c))
	.cloned(),
	);
	Ok(())
	})?;
	if missing_cols.is_empty() {
	return Ok(Self::new(LogicalPlan::Sort(Sort {
	expr: normalize_sorts(sorts, &self.plan)?,
	input: self.plan,
	fetch,
	})));
	}

We have count(*) (alias) and count(1) (column name) which mismatches in missing_cols in sort.

[jayzhan211]

#14895

[jayzhan211]

why empty project?

[jayzhan211]

optimize_projections remove exprs in projection but keep the empty one

[2025-02-28T06:22:26Z DEBUG datafusion_optimizer::utils] simplify_expressions:
    Projection: t1.a
      LeftSemi Join: 
        Filter: Boolean(true)
          TableScan: t1
        SubqueryAlias: __correlated_sq_1
          Projection: count(Int64(1)) AS count(*)
            Aggregate: groupBy=[[]], aggr=[[count(Int64(1))]]
              TableScan: t2
    
[2025-02-28T06:22:26Z DEBUG datafusion_optimizer::optimizer] Plan unchanged by optimizer rule 'unwrap_cast_in_comparison' (pass 0)
[2025-02-28T06:22:26Z DEBUG datafusion_optimizer::optimizer] Plan unchanged by optimizer rule 'common_sub_expression_eliminate' (pass 0)
[2025-02-28T06:22:26Z DEBUG datafusion_optimizer::optimizer] Plan unchanged by optimizer rule 'eliminate_group_by_constant' (pass 0)
[2025-02-28T06:22:26Z DEBUG datafusion_optimizer::utils] optimize_projections:
    LeftSemi Join: 
      Filter: Boolean(true)
        TableScan: t1 projection=[a]
      SubqueryAlias: __correlated_sq_1
        Projection: 
          Aggregate: groupBy=[[]], aggr=[[count(Int64(1))]]
            TableScan: t2 projection=[]

[alamb]

weird

[alamb]

Thanks @jayzhan211 -- I am not sure about the column name functions, but otherwise this is looking very nice 👍

[alamb]

I don't really understand what this function is for. It seems pretty confusing

Maye we could add a doc example or something to make it less confusing?

Likewise for the count_all_window_column function

[jayzhan211]

One of the example is been used in test_count_wildcard_on_where_scalar_subquery

 let df_results = ctx
        .table("t1")
        .await?
        .filter(
            scalar_subquery(Arc::new(
                ctx.table("t2")
                    .await?
                    .filter(out_ref_col(DataType::UInt32, "t1.a").eq(col("t2.a")))?
                    .aggregate(vec![], vec![count_all()])?
                    .select(vec![count_all_column()])?
                    .into_unoptimized_plan(),
            ))
            .gt(lit(ScalarValue::UInt8(Some(0)))),
        )?
        .select(vec![col("t1.a"), col("t1.b")])?
        .explain(false, false)?
        .collect()
        .await?;

[alamb]

weird