Use struct instead of named_struct when there are no aliases (apache#9897)

* Revert "use alias (apache#9894)"

This reverts commit 9487ca0.

* Use `struct` instead of `named_struct` when there are no aliases

* Update docs

* fmt

Author: alamb

datafusion/sql/src/expr/mod.rs

@@ -593,6 +593,7 @@ impl<'a, S: ContextProvider> SqlToRel<'a, S> {
         }
     }
 
+    /// Parses a struct(..) expression
     fn parse_struct(
         &self,
         values: Vec<SQLExpr>,
@@ -603,6 +604,25 @@ impl<'a, S: ContextProvider> SqlToRel<'a, S> {
         if !fields.is_empty() {
             return not_impl_err!("Struct fields are not supported yet");
         }
+
+        if values
+            .iter()
+            .any(|value| matches!(value, SQLExpr::Named { .. }))
+        {
+            self.create_named_struct(values, input_schema, planner_context)
+        } else {
+            self.create_struct(values, input_schema, planner_context)
+        }
+    }
+
+    // Handles a call to struct(...) where the arguments are named. For example
+    // `struct (v as foo, v2 as bar)` by creating a call to the `named_struct` function
+    fn create_named_struct(
+        &self,
+        values: Vec<SQLExpr>,
+        input_schema: &DFSchema,
+        planner_context: &mut PlannerContext,
+    ) -> Result<Expr> {
         let args = values
             .into_iter()
             .enumerate()
@@ -647,6 +667,34 @@ impl<'a, S: ContextProvider> SqlToRel<'a, S> {
         )))
     }
 
+    // Handles a call to struct(...) where the arguments are not named. For example
+    // `struct (v, v2)` by creating a call to the `struct` function
+    // which will create a struct with fields named `c0`, `c1`, etc.
+    fn create_struct(
+        &self,
+        values: Vec<SQLExpr>,
+        input_schema: &DFSchema,
+        planner_context: &mut PlannerContext,
+    ) -> Result<Expr> {
+        let args = values
+            .into_iter()
+            .map(|value| {
+                self.sql_expr_to_logical_expr(value, input_schema, planner_context)
+            })
+            .collect::<Result<Vec<_>>>()?;
+        let struct_func = self
+            .context_provider
+            .get_function_meta("struct")
+            .ok_or_else(|| {
+                internal_datafusion_err!("Unable to find expected 'struct' function")
+            })?;
+
+        Ok(Expr::ScalarFunction(ScalarFunction::new_udf(
+            struct_func,
+            args,
+        )))
+    }
+
     fn parse_array_agg(
         &self,
         array_agg: ArrayAgg,

datafusion/sqllogictest/test_files/explain.slt

@@ -390,8 +390,8 @@ query TT
 explain select struct(1, 2.3, 'abc');
 ----
 logical_plan
-Projection: Struct({c0:1,c1:2.3,c2:abc}) AS named_struct(Utf8("c0"),Int64(1),Utf8("c1"),Float64(2.3),Utf8("c2"),Utf8("abc"))
+Projection: Struct({c0:1,c1:2.3,c2:abc}) AS struct(Int64(1),Float64(2.3),Utf8("abc"))
 --EmptyRelation
 physical_plan
-ProjectionExec: expr=[{c0:1,c1:2.3,c2:abc} as named_struct(Utf8("c0"),Int64(1),Utf8("c1"),Float64(2.3),Utf8("c2"),Utf8("abc"))]
+ProjectionExec: expr=[{c0:1,c1:2.3,c2:abc} as struct(Int64(1),Float64(2.3),Utf8("abc"))]
 --PlaceholderRowExec

datafusion/sqllogictest/test_files/struct.slt

@@ -85,10 +85,10 @@ query TT
 explain select struct(a, b, c) from values;
 ----
 logical_plan
-Projection: named_struct(Utf8("c0"), values.a, Utf8("c1"), values.b, Utf8("c2"), values.c)
+Projection: struct(values.a, values.b, values.c)
 --TableScan: values projection=[a, b, c]
 physical_plan
-ProjectionExec: expr=[named_struct(c0, a@0, c1, b@1, c2, c@2) as named_struct(Utf8("c0"),values.a,Utf8("c1"),values.b,Utf8("c2"),values.c)]
+ProjectionExec: expr=[struct(a@0, b@1, c@2) as struct(values.a,values.b,values.c)]
 --MemoryExec: partitions=1, partition_sizes=[1]
 
 # error on 0 arguments
@@ -179,4 +179,4 @@ drop table values;
 query T
 select arrow_typeof(named_struct('first', 1, 'second', 2, 'third', 3));
 ----
-Struct([Field { name: "first", data_type: Int64, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {} }, Field { name: "second", data_type: Int64, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {} }, Field { name: "third", data_type: Int64, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {} }])
+Struct([Field { name: "first", data_type: Int64, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {} }, Field { name: "second", data_type: Int64, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {} }, Field { name: "third", data_type: Int64, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {} }])

docs/source/user-guide/sql/scalar_functions.md

@@ -3336,6 +3336,16 @@ select * from t;
 | 3 | 4 |
 +---+---+
 
+-- use default names `c0`, `c1`
+❯ select struct(a, b) from t;
++-----------------+
+| struct(t.a,t.b) |
++-----------------+
+| {c0: 1, c1: 2}  |
+| {c0: 3, c1: 4}  |
++-----------------+
+
+-- name the first field `field_a`
 select struct(a as field_a, b) from t;
 +--------------------------------------------------+
 | named_struct(Utf8("field_a"),t.a,Utf8("c1"),t.b) |