Merge pull request #37667 from ClickHouse/group-by-enum-fix

Support types with non-standard defaults in ROLLUP, CUBE, GROUPING SETS

src/Columns/IColumn.h

@@ -90,7 +90,7 @@ public:
     /// Creates column with the same type and specified size.
     /// If size is less current size, then data is cut.
     /// If size is greater, than default values are appended.
-    [[nodiscard]] virtual MutablePtr cloneResized(size_t /*size*/) const { throw Exception("Cannot cloneResized() column " + getName(), ErrorCodes::NOT_IMPLEMENTED); }
+    [[nodiscard]] virtual MutablePtr cloneResized(size_t /*size*/) const { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Cannot cloneResized() column {}", getName()); }
 
     /// Returns number of values in column.
     [[nodiscard]] virtual size_t size() const = 0;

src/DataTypes/IDataType.cpp

@@ -1,3 +1,4 @@
+#include <cstddef>
 #include <Columns/IColumn.h>
 #include <Columns/ColumnConst.h>
 #include <Columns/ColumnSparse.h>
@@ -162,6 +163,12 @@ void IDataType::insertDefaultInto(IColumn & column) const
     column.insertDefault();
 }
 
+void IDataType::insertManyDefaultsInto(IColumn & column, size_t n) const
+{
+    for (size_t i = 0; i < n; ++i)
+        insertDefaultInto(column);
+}
+
 void IDataType::setCustomization(DataTypeCustomDescPtr custom_desc_) const
 {
     /// replace only if not null

src/DataTypes/IDataType.h

@@ -159,6 +159,8 @@ public:
       */
     virtual void insertDefaultInto(IColumn & column) const;
 
+    void insertManyDefaultsInto(IColumn & column, size_t n) const;
+
     /// Checks that two instances belong to the same type
     virtual bool equals(const IDataType & rhs) const = 0;
 

src/Processors/QueryPlan/AggregatingStep.cpp

@@ -241,7 +241,9 @@ void AggregatingStep::transformPipeline(QueryPipelineBuilder & pipeline, const B
                     if (missign_column_index < missing_columns.size() && missing_columns[missign_column_index] == i)
                     {
                         ++missign_column_index;
-                        auto column = ColumnConst::create(col.column->cloneResized(1), 0);
+                        auto column_with_default = col.column->cloneEmpty();
+                        col.type->insertDefaultInto(*column_with_default);
+                        auto column = ColumnConst::create(std::move(column_with_default), 0);
                         const auto * node = &dag->addColumn({ColumnPtr(std::move(column)), col.type, col.name});
                         node = &dag->materializeNode(*node);
                         index.push_back(node);

src/Processors/Transforms/CubeTransform.cpp

@@ -35,6 +35,8 @@ void CubeTransform::consume(Chunk chunk)
     consumed_chunks.emplace_back(std::move(chunk));
 }
 
+MutableColumnPtr getColumnWithDefaults(Block const & header, size_t key, size_t n);
+
 Chunk CubeTransform::generate()
 {
     if (!consumed_chunks.empty())
@@ -53,8 +55,9 @@ Chunk CubeTransform::generate()
         current_zero_columns.clear();
         current_zero_columns.reserve(keys.size());
 
+        auto const & input_header = getInputPort().getHeader();
         for (auto key : keys)
-            current_zero_columns.emplace_back(current_columns[key]->cloneEmpty()->cloneResized(num_rows));
+            current_zero_columns.emplace_back(getColumnWithDefaults(input_header, key, num_rows));
     }
 
     auto gen_chunk = std::move(cube_chunk);

src/Processors/Transforms/RollupTransform.cpp

@@ -29,6 +29,14 @@ Chunk RollupTransform::merge(Chunks && chunks, bool final)
     return Chunk(rollup_block.getColumns(), num_rows);
 }
 
+MutableColumnPtr getColumnWithDefaults(Block const & header, size_t key, size_t n)
+{
+    auto const & col = header.getByPosition(key);
+    auto result_column = col.column->cloneEmpty();
+    col.type->insertManyDefaultsInto(*result_column, n);
+    return result_column;
+}
+
 Chunk RollupTransform::generate()
 {
     if (!consumed_chunks.empty())
@@ -51,7 +59,7 @@ Chunk RollupTransform::generate()
 
         auto num_rows = gen_chunk.getNumRows();
         auto columns = gen_chunk.getColumns();
-        columns[key] = columns[key]->cloneEmpty()->cloneResized(num_rows);
+        columns[key] = getColumnWithDefaults(getInputPort().getHeader(), key, num_rows);
 
         Chunks chunks;
         chunks.emplace_back(std::move(columns), num_rows);

tests/queries/0_stateless/02313_group_by_modifiers_with_non-default_types.reference

@@ -0,0 +1,113 @@
+-- { echoOn }
+SELECT
+    count() as d, a, b, c
+FROM test02313
+GROUP BY ROLLUP(a, b, c)
+ORDER BY d, a, b, c;
+1	one	default	0
+1	one	default	2
+1	one	default	4
+1	one	default	6
+1	one	default	8
+1	two	non-default	1
+1	two	non-default	3
+1	two	non-default	5
+1	two	non-default	7
+1	two	non-default	9
+5	one	default	0
+5	one	default	0
+5	two	default	0
+5	two	non-default	0
+10	one	default	0
+SELECT
+    count() as d, a, b, c
+FROM test02313
+GROUP BY CUBE(a, b, c)
+ORDER BY d, a, b, c;
+1	one	default	0
+1	one	default	0
+1	one	default	0
+1	one	default	0
+1	one	default	1
+1	one	default	2
+1	one	default	2
+1	one	default	2
+1	one	default	2
+1	one	default	3
+1	one	default	4
+1	one	default	4
+1	one	default	4
+1	one	default	4
+1	one	default	5
+1	one	default	6
+1	one	default	6
+1	one	default	6
+1	one	default	6
+1	one	default	7
+1	one	default	8
+1	one	default	8
+1	one	default	8
+1	one	default	8
+1	one	default	9
+1	one	non-default	1
+1	one	non-default	3
+1	one	non-default	5
+1	one	non-default	7
+1	one	non-default	9
+1	two	default	1
+1	two	default	3
+1	two	default	5
+1	two	default	7
+1	two	default	9
+1	two	non-default	1
+1	two	non-default	3
+1	two	non-default	5
+1	two	non-default	7
+1	two	non-default	9
+5	one	default	0
+5	one	default	0
+5	one	default	0
+5	one	non-default	0
+5	two	default	0
+5	two	non-default	0
+10	one	default	0
+SELECT
+    count() as d, a, b, c
+FROM test02313
+GROUP BY GROUPING SETS
+    (
+        (c),
+        (a, c),
+        (b, c)
+    )
+ORDER BY d, a, b, c;
+1	one	default	0
+1	one	default	0
+1	one	default	0
+1	one	default	1
+1	one	default	2
+1	one	default	2
+1	one	default	2
+1	one	default	3
+1	one	default	4
+1	one	default	4
+1	one	default	4
+1	one	default	5
+1	one	default	6
+1	one	default	6
+1	one	default	6
+1	one	default	7
+1	one	default	8
+1	one	default	8
+1	one	default	8
+1	one	default	9
+1	one	non-default	1
+1	one	non-default	3
+1	one	non-default	5
+1	one	non-default	7
+1	one	non-default	9
+1	two	default	1
+1	two	default	3
+1	two	default	5
+1	two	default	7
+1	two	default	9

tests/queries/0_stateless/02313_group_by_modifiers_with_non-default_types.sql

@@ -0,0 +1,39 @@
+DROP TABLE IF EXISTS test02313;
+
+CREATE TABLE test02313
+(
+    a Enum('one' = 1, 'two' = 2),
+    b Enum('default' = 0, 'non-default' = 1),
+    c UInt8
+)
+ENGINE = MergeTree()
+ORDER BY (a, b, c);
+
+INSERT INTO test02313 SELECT number % 2 + 1 AS a, number % 2 AS b, number FROM numbers(10);
+
+-- { echoOn }
+SELECT
+    count() as d, a, b, c
+FROM test02313
+GROUP BY ROLLUP(a, b, c)
+ORDER BY d, a, b, c;
+
+SELECT
+    count() as d, a, b, c
+FROM test02313
+GROUP BY CUBE(a, b, c)
+ORDER BY d, a, b, c;
+
+SELECT
+    count() as d, a, b, c
+FROM test02313
+GROUP BY GROUPING SETS
+    (
+        (c),
+        (a, c),
+        (b, c)
+    )
+ORDER BY d, a, b, c;
+
+-- { echoOff }
+DROP TABLE test02313;