feat(optimizer)!: Alias expanded USING STRUCT fields (#4474)

* feat(optimizer): Alias expanded USING STRUCT fields

* Typo

Author: VaggelisD

sqlglot/optimizer/qualify_columns.py

@@ -220,11 +220,16 @@ def _update_source_columns(source_name: str) -> None:
             if not column.table and column.name in column_tables:
                 tables = column_tables[column.name]
                 coalesce_args = [exp.column(column.name, table=table) for table in tables]
-                replacement = exp.func("coalesce", *coalesce_args)
+                replacement: exp.Expression = exp.func("coalesce", *coalesce_args)
 
-                # Ensure selects keep their output name
                 if isinstance(column.parent, exp.Select):
+                    # Ensure the USING column keeps its name if it's projected
                     replacement = alias(replacement, alias=column.name, copy=False)
+                elif isinstance(column.parent, exp.Struct):
+                    # Ensure the USING column keeps its name if it's an anonymous STRUCT field
+                    replacement = exp.PropertyEQ(
+                        this=exp.to_identifier(column.name), expression=replacement
+                    )
 
                 scope.replace(column, replacement)
 

tests/fixtures/optimizer/qualify_columns.sql

@@ -532,6 +532,14 @@ WITH t1 AS (SELECT 'x' AS id, CAST('2024-01-01' AS DATE) AS foo, 000 AS value),
 WITH t1 AS (SELECT 'x' AS id, CAST('2024-01-01' AS DATE) AS foo, 000 AS value), t2 AS (SELECT 'x' AS id, CAST('2024-02-02' AS DATE) AS foo, 123 AS value), t3 AS (SELECT 'x' AS id, CAST('2024-02-02' AS DATE) AS foo, 456 AS value), t4 AS (SELECT 'x' AS id, CAST('2024-03-03' AS DATE) AS foo, 789 AS value) SELECT * FROM t1 FULL OUTER JOIN t2 USING (id, foo) FULL OUTER JOIN t3 USING (id, foo) FULL OUTER JOIN t4 USING (id, foo);
 WITH t1 AS (SELECT 'x' AS id, CAST('2024-01-01' AS DATE) AS foo, 000 AS value), t2 AS (SELECT 'x' AS id, CAST('2024-02-02' AS DATE) AS foo, 123 AS value), t3 AS (SELECT 'x' AS id, CAST('2024-02-02' AS DATE) AS foo, 456 AS value), t4 AS (SELECT 'x' AS id, CAST('2024-03-03' AS DATE) AS foo, 789 AS value) SELECT COALESCE(t1.id, t2.id, t3.id, t4.id) AS id, COALESCE(t1.foo, t2.foo, t3.foo, t4.foo) AS foo, t1.value AS value, t2.value AS value, t3.value AS value, t4.value AS value FROM t1 AS t1 FULL OUTER JOIN t2 AS t2 ON t1.id = t2.id AND t1.foo = t2.foo FULL OUTER JOIN t3 AS t3 ON COALESCE(t1.id, t2.id) = t3.id AND COALESCE(t1.foo, t2.foo) = t3.foo FULL OUTER JOIN t4 AS t4 ON COALESCE(t1.id, t2.id, t3.id) = t4.id AND COALESCE(t1.foo, t2.foo, t3.foo) = t4.foo;
 
+# title: Name anonymous STRUCT fields if replacing USING columns
+WITH t1 AS (SELECT 1 AS id), t2 AS (SELECT 2 AS id) SELECT STRUCT(id) AS my_field FROM t1 JOIN t2 USING (id);
+WITH t1 AS (SELECT 1 AS id), t2 AS (SELECT 2 AS id) SELECT STRUCT(COALESCE(t1.id, t2.id) AS id) AS my_field FROM t1 AS t1 JOIN t2 AS t2 ON t1.id = t2.id;
+
+# title: Do not rename aliased STRUCT fields if replacing USING columns
+WITH t1 AS (SELECT 1 AS id), t2 AS (SELECT 2 AS id) SELECT STRUCT(id AS col) AS my_field FROM t1 JOIN t2 USING (id);
+WITH t1 AS (SELECT 1 AS id), t2 AS (SELECT 2 AS id) SELECT STRUCT(COALESCE(t1.id, t2.id) AS col) AS my_field FROM t1 AS t1 JOIN t2 AS t2 ON t1.id = t2.id;
+
 --------------------------------------
 -- Hint with table reference
 --------------------------------------