Merge pull request duckdb#9626 from Mytherin/pivotparameter

Fix duckdb#9548 - Throw a more clear error when using parameters inside of the source of a top-level PIVOT statement

src/include/duckdb/parser/transformer.hpp

@@ -48,6 +48,7 @@ class Transformer {
 		unique_ptr<SelectNode> base;
 		unique_ptr<ParsedExpression> column;
 		unique_ptr<QueryNode> subquery;
+		bool has_parameters;
 	};
 
 public:
@@ -90,7 +91,7 @@ class Transformer {
 	bool GetParam(const string &name, idx_t &index, PreparedParamType type);
 
 	void AddPivotEntry(string enum_name, unique_ptr<SelectNode> source, unique_ptr<ParsedExpression> column,
-	                   unique_ptr<QueryNode> subquery);
+	                   unique_ptr<QueryNode> subquery, bool has_parameters);
 	unique_ptr<SQLStatement> GenerateCreateEnumStmt(unique_ptr<CreatePivotEntry> entry);
 	bool HasPivotEntries();
 	idx_t PivotEntryCount();

src/parser/transform/statement/transform_pivot_stmt.cpp

@@ -19,16 +19,18 @@
 namespace duckdb {
 
 void Transformer::AddPivotEntry(string enum_name, unique_ptr<SelectNode> base, unique_ptr<ParsedExpression> column,
-                                unique_ptr<QueryNode> subquery) {
+                                unique_ptr<QueryNode> subquery, bool has_parameters) {
 	if (parent) {
-		parent->AddPivotEntry(std::move(enum_name), std::move(base), std::move(column), std::move(subquery));
+		parent->AddPivotEntry(std::move(enum_name), std::move(base), std::move(column), std::move(subquery),
+		                      has_parameters);
 		return;
 	}
 	auto result = make_uniq<CreatePivotEntry>();
 	result->enum_name = std::move(enum_name);
 	result->base = std::move(base);
 	result->column = std::move(column);
 	result->subquery = std::move(subquery);
+	result->has_parameters = has_parameters;
 
 	pivot_entries.push_back(std::move(result));
 }
@@ -113,6 +115,13 @@ unique_ptr<SQLStatement> Transformer::GenerateCreateEnumStmt(unique_ptr<CreatePi
 unique_ptr<SQLStatement> Transformer::CreatePivotStatement(unique_ptr<SQLStatement> statement) {
 	auto result = make_uniq<MultiStatement>();
 	for (auto &pivot : pivot_entries) {
+		if (pivot->has_parameters) {
+			throw ParserException(
+			    "PIVOT statements with pivot elements extracted from the data cannot have parameters in their source.\n"
+			    "In order to use parameters the PIVOT values must be manually specified, e.g.:\n"
+			    "PIVOT ... ON %s IN (val1, val2, ...)",
+			    pivot->column->ToString());
+		}
 		result->statements.push_back(GenerateCreateEnumStmt(std::move(pivot)));
 	}
 	result->statements.push_back(std::move(statement));
@@ -125,7 +134,10 @@ unique_ptr<SQLStatement> Transformer::CreatePivotStatement(unique_ptr<SQLStateme
 
 unique_ptr<QueryNode> Transformer::TransformPivotStatement(duckdb_libpgquery::PGSelectStmt &select) {
 	auto pivot = select.pivot;
+	auto current_param_count = ParamCount();
 	auto source = TransformTableRefNode(*pivot->source);
+	auto next_param_count = ParamCount();
+	bool has_parameters = next_param_count > current_param_count;
 
 	auto select_node = make_uniq<SelectNode>();
 	vector<unique_ptr<CTENode>> materialized_ctes;
@@ -171,7 +183,8 @@ unique_ptr<QueryNode> Transformer::TransformPivotStatement(duckdb_libpgquery::PG
 		auto new_select = make_uniq<SelectNode>();
 		ExtractCTEsRecursive(new_select->cte_map);
 		new_select->from_table = source->Copy();
-		AddPivotEntry(enum_name, std::move(new_select), col.pivot_expressions[0]->Copy(), std::move(col.subquery));
+		AddPivotEntry(enum_name, std::move(new_select), col.pivot_expressions[0]->Copy(), std::move(col.subquery),
+		              has_parameters);
 		col.pivot_enum = enum_name;
 	}
 

test/sql/pivot/pivot_prepare.test

@@ -7,27 +7,27 @@ CREATE OR REPLACE TABLE monthly_sales(empid INT, amount INT, month TEXT);
 
 statement ok
 INSERT INTO monthly_sales VALUES
-    (1, 10000, 'JAN'),
-    (1, 400, 'JAN'),
-    (2, 4500, 'JAN'),
-    (2, 35000, 'JAN'),
-    (1, 5000, 'FEB'),
-    (1, 3000, 'FEB'),
-    (2, 200, 'FEB'),
-    (2, 90500, 'FEB'),
-    (1, 6000, 'MAR'),
-    (1, 5000, 'MAR'),
-    (2, 2500, 'MAR'),
-    (2, 9500, 'MAR'),
-    (1, 8000, 'APR'),
-    (1, 10000, 'APR'),
-    (2, 800, 'APR'),
-    (2, 4500, 'APR');
+    (1, 10000, '1-JAN'),
+    (1, 400, '1-JAN'),
+    (2, 4500, '1-JAN'),
+    (2, 35000, '1-JAN'),
+    (1, 5000, '2-FEB'),
+    (1, 3000, '2-FEB'),
+    (2, 200, '2-FEB'),
+    (2, 90500, '2-FEB'),
+    (1, 6000, '3-MAR'),
+    (1, 5000, '3-MAR'),
+    (2, 2500, '3-MAR'),
+    (2, 9500, '3-MAR'),
+    (1, 8000, '4-APR'),
+    (1, 10000, '4-APR'),
+    (2, 800, '4-APR'),
+    (2, 4500, '4-APR');
 
 statement ok
 PREPARE v1 AS SELECT *
   FROM monthly_sales
-    PIVOT(SUM(amount + ?) FOR MONTH IN ('JAN', 'FEB', 'MAR', 'APR'))
+    PIVOT(SUM(amount + ?) FOR MONTH IN ('1-JAN', '2-FEB', '3-MAR', '4-APR'))
       AS p
   ORDER BY EMPID;
 
@@ -43,3 +43,21 @@ EXECUTE v1(1)
 ----
 1	10402	8002	11002	18002
 2	39502	90702	12002	5302
+
+# prepare top-level pivot stmt
+statement ok
+PREPARE v2 AS
+   PIVOT monthly_sales ON MONTH USING SUM(AMOUNT + ?)
+
+query IIIII rowsort
+EXECUTE v2(1)
+----
+1	10402	8002	11002	18002
+2	39502	90702	12002	5302
+
+# parameters within subquery of top-level pivot statement not supported
+statement error
+PREPARE v3 AS
+   PIVOT (SELECT empid, amount + ? AS amount, month FROM monthly_sales) ON MONTH USING SUM(AMOUNT)
+----
+cannot have parameters in their source