Merge pull request #254 from den818/MapColumn

Support column type: map

Author: Enmk

README.md

@@ -21,6 +21,7 @@ C++ client for [ClickHouse](https://clickhouse.com/).
 * UInt8, UInt16, UInt32, UInt64, Int8, Int16, Int32, Int64
 * Int128
 * UUID
+* Map
 
 ## Building
 
@@ -87,7 +88,7 @@ int main()
 
 ## Retries
 If you wish to implement some retry logic atop of `clickhouse::Client` there are few simple rules to make you life easier:
-- If previous attempt threw an exception, then make sure to call `clickhouse::Client::ResetConnection()` before the next try. 
+- If previous attempt threw an exception, then make sure to call `clickhouse::Client::ResetConnection()` before the next try.
 - For `clickhouse::Client::Insert()` you can reuse a block from previous try, no need to rebuild it from scratch.
 
 See https://github.com/ClickHouse/clickhouse-cpp/issues/184 for details.

clickhouse/CMakeLists.txt

@@ -18,6 +18,7 @@ SET ( clickhouse-cpp-lib-src
     columns/lowcardinalityadaptor.h
     columns/nullable.cpp
     columns/numeric.cpp
+    columns/map.cpp
     columns/string.cpp
     columns/tuple.cpp
     columns/uuid.cpp
@@ -121,6 +122,7 @@ INSTALL(FILES columns/itemview.h DESTINATION include/clickhouse/columns/)
 INSTALL(FILES columns/lowcardinality.h DESTINATION include/clickhouse/columns/)
 INSTALL(FILES columns/nullable.h DESTINATION include/clickhouse/columns/)
 INSTALL(FILES columns/numeric.h DESTINATION include/clickhouse/columns/)
+INSTALL(FILES columns/map.h DESTINATION include/clickhouse/columns/)
 INSTALL(FILES columns/string.h DESTINATION include/clickhouse/columns/)
 INSTALL(FILES columns/tuple.h DESTINATION include/clickhouse/columns/)
 INSTALL(FILES columns/utils.h DESTINATION include/clickhouse/columns/)

clickhouse/client.h

@@ -12,6 +12,7 @@
 #include "columns/lowcardinality.h"
 #include "columns/nullable.h"
 #include "columns/numeric.h"
+#include "columns/map.h"
 #include "columns/string.h"
 #include "columns/tuple.h"
 #include "columns/uuid.h"

clickhouse/columns/array.h

@@ -2,6 +2,7 @@
 
 #include "column.h"
 #include "numeric.h"
+#include "utils.h"
 
 #include <memory>
 
@@ -121,13 +122,8 @@ class ColumnArrayT : public ColumnArray {
      *  This is a static method to make such conversion verbose.
      */
     static auto Wrap(ColumnArray&& col) {
-        if constexpr (std::is_base_of_v<ColumnArray, NestedColumnType> && !std::is_same_v<ColumnArray, NestedColumnType>) {
-            // assuming NestedColumnType is ArrayT specialization
-            return std::make_shared<ColumnArrayT<NestedColumnType>>(NestedColumnType::Wrap(col.GetData()), col.offsets_);
-        } else {
-            auto nested_data = col.GetData()->template AsStrict<NestedColumnType>();
-            return std::make_shared<ColumnArrayT<NestedColumnType>>(nested_data, col.offsets_);
-        }
+        auto nested_data = WrapColumn<NestedColumnType>(col.GetData());
+        return std::make_shared<ColumnArrayT<NestedColumnType>>(nested_data, col.offsets_);
     }
 
     static auto Wrap(Column&& col) {
@@ -146,7 +142,7 @@ class ColumnArrayT : public ColumnArray {
         const size_t size_;
 
     public:
-        using ValueType = typename NestedColumnType::ValueType;
+        using ValueType = std::decay_t<decltype(std::declval<NestedColumnType>().At(0))>;
 
         ArrayValueView(std::shared_ptr<NestedColumnType> data, size_t offset = 0, size_t size = std::numeric_limits<size_t>::max())
             : typed_nested_data_(data)
@@ -178,7 +174,7 @@ class ColumnArrayT : public ColumnArray {
                 , index_(index)
             {}
 
-            using ValueType = typename NestedColumnType::ValueType;
+            using ValueType = typename ArrayValueView::ValueType;
 
             inline auto operator*() const {
                 return typed_nested_data_->At(offset_ + index_);
@@ -226,6 +222,22 @@ class ColumnArrayT : public ColumnArray {
         inline size_t Size() const {
             return size_;
         }
+
+        inline bool operator==(const ArrayValueView& other) const {
+            if (size() != other.size()) {
+                return false;
+            }
+            for (size_t i = 0; i < size_; ++i) {
+                if ((*this)[i] != other[i]) {
+                    return false;
+                }
+            }
+            return true;
+        }
+
+        inline bool operator!=(const ArrayValueView& other) const {
+            return !(*this == other);
+        }
     };
 
     inline auto At(size_t index) const {
@@ -267,6 +279,20 @@ class ColumnArrayT : public ColumnArray {
         AddOffset(counter);
     }
 
+    ColumnRef Slice(size_t begin, size_t size) const override {
+        return Wrap(ColumnArray::Slice(begin, size));
+    }
+
+    ColumnRef CloneEmpty() const override {
+        return Wrap(ColumnArray::CloneEmpty());
+    }
+
+    void Swap(Column& other) override {
+        auto & col = dynamic_cast<ColumnArrayT<NestedColumnType> &>(other);
+        typed_nested_data_.swap(col.typed_nested_data_);
+        ColumnArray::Swap(other);
+    }
+
 private:
     /// Helper to allow wrapping a "typeless" ColumnArray
     ColumnArrayT(ColumnArray&& array, std::shared_ptr<NestedColumnType> nested_data)

clickhouse/columns/factory.cpp

@@ -14,6 +14,7 @@
 #include "string.h"
 #include "tuple.h"
 #include "uuid.h"
+#include "map.h"
 
 #include "../types/type_parser.h"
 
@@ -86,7 +87,7 @@ static ColumnRef CreateTerminalColumn(const TypeAst& ast) {
     case Type::Date:
         return std::make_shared<ColumnDate>();
     case Type::Date32:
-        return std::make_shared<ColumnDate32>();    
+        return std::make_shared<ColumnDate32>();
 
     case Type::IPv4:
         return std::make_shared<ColumnIPv4>();
@@ -198,6 +199,27 @@ static ColumnRef CreateColumnFromAst(const TypeAst& ast, CreateColumnByTypeSetti
             return CreateTerminalColumn(ast.elements.back());
         }
 
+        case TypeAst::Map: {
+            if (ast.elements.size() != 2) {
+                throw ValidationError(ast.name + " content is not correct");
+            }
+
+            std::vector<ColumnRef> columns;
+
+            columns.reserve(ast.elements.size());
+            for (const auto& elem : ast.elements) {
+                if (auto col = CreateColumnFromAst(elem, settings)) {
+                    columns.push_back(col);
+                } else {
+                    return nullptr;
+                }
+            }
+
+            return std::make_shared<ColumnMap>(
+                std::make_shared<ColumnArray>(
+                    std::make_shared<ColumnTuple>(columns)));
+        }
+
         case TypeAst::Assign:
         case TypeAst::Null:
         case TypeAst::Number:

clickhouse/columns/itemview.cpp

@@ -77,6 +77,7 @@ void ItemView::ValidateData(Type::Code type, DataType data) {
         case Type::Code::Nullable:
         case Type::Code::Tuple:
         case Type::Code::LowCardinality:
+        case Type::Code::Map:
             throw AssertionError("Unsupported type in ItemView: " + std::string(Type::TypeName(type)));
 
         case Type::Code::IPv6:

clickhouse/columns/map.cpp

@@ -0,0 +1,83 @@
+#include "map.h"
+
+#include <stdexcept>
+
+#include "../exceptions.h"
+#include "utils.h"
+
+namespace {
+
+using namespace clickhouse;
+
+TypeRef GetMapType(const Type& data_type) {
+    auto array = data_type.As<ArrayType>();
+    if (!array) {
+        throw ValidationError("Wrong type  " + data_type.GetName() + " of data for map");
+    }
+    auto tuple = array->GetItemType()->As<TupleType>();
+    if (!tuple) {
+        throw ValidationError("Wrong type  " + data_type.GetName() + " of data for map");
+    }
+    auto types = tuple->GetTupleType();
+    if (types.size() != 2) {
+        throw ValidationError("Wrong type  " + data_type.GetName() + " of data for map");
+    }
+    return Type::CreateMap(types[0], types[1]);
+}
+
+}  // namespace
+
+namespace clickhouse {
+
+ColumnMap::ColumnMap(ColumnRef data)
+    : Column(GetMapType(data->GetType())), data_(data->As<ColumnArray>()) {
+}
+
+void ColumnMap::Clear() {
+    data_->Clear();
+}
+
+void ColumnMap::Append(ColumnRef column) {
+    if (auto col = column->As<ColumnMap>()) {
+        data_->Append(col->data_);
+    }
+}
+
+bool ColumnMap::LoadPrefix(InputStream* input, size_t rows) {
+    return data_->LoadPrefix(input, rows);
+}
+
+bool ColumnMap::LoadBody(InputStream* input, size_t rows) {
+    return data_->LoadBody(input, rows);
+}
+
+void ColumnMap::SavePrefix(OutputStream* output) {
+    data_->SavePrefix(output);
+}
+
+void ColumnMap::SaveBody(OutputStream* output) {
+    data_->SaveBody(output);
+}
+
+size_t ColumnMap::Size() const {
+    return data_->Size();
+}
+
+ColumnRef ColumnMap::Slice(size_t begin, size_t len) const {
+    return std::make_shared<ColumnMap>(data_->Slice(begin, len));
+}
+
+ColumnRef ColumnMap::CloneEmpty() const {
+    return std::make_shared<ColumnMap>(data_->CloneEmpty());
+}
+
+void ColumnMap::Swap(Column& other) {
+    auto& col = dynamic_cast<ColumnMap&>(other);
+    data_.swap(col.data_);
+}
+
+ColumnRef ColumnMap::GetAsColumn(size_t n) const {
+    return data_->GetAsColumn(n);
+}
+
+}  // namespace clickhouse