| 1 | #pragma once |
| 2 | |
| 3 | #include <Columns/IColumn.h> |
| 4 | #include <Columns/IColumnImpl.h> |
| 5 | #include <Columns/ColumnsNumber.h> |
| 6 | #include <Common/typeid_cast.h> |
| 7 | #include <Common/assert_cast.h> |
| 8 | |
| 9 | |
| 10 | namespace DB |
| 11 | { |
| 12 | |
| 13 | using NullMap = ColumnUInt8::Container; |
| 14 | using ConstNullMapPtr = const NullMap *; |
| 15 | |
| 16 | /// Class that specifies nullable columns. A nullable column represents |
| 17 | /// a column, which may have any type, provided with the possibility of |
| 18 | /// storing NULL values. For this purpose, a ColumNullable object stores |
| 19 | /// an ordinary column along with a special column, namely a byte map, |
| 20 | /// whose type is ColumnUInt8. The latter column indicates whether the |
| 21 | /// value of a given row is a NULL or not. Such a design is preferred |
| 22 | /// over a bitmap because columns are usually stored on disk as compressed |
| 23 | /// files. In this regard, using a bitmap instead of a byte map would |
| 24 | /// greatly complicate the implementation with little to no benefits. |
| 25 | class ColumnNullable final : public COWHelper<IColumn, ColumnNullable> |
| 26 | { |
| 27 | private: |
| 28 | friend class COWHelper<IColumn, ColumnNullable>; |
| 29 | |
| 30 | ColumnNullable(MutableColumnPtr && nested_column_, MutableColumnPtr && null_map_); |
| 31 | ColumnNullable(const ColumnNullable &) = default; |
| 32 | |
| 33 | public: |
| 34 | /** Create immutable column using immutable arguments. This arguments may be shared with other columns. |
| 35 | * Use IColumn::mutate in order to make mutable column and mutate shared nested columns. |
| 36 | */ |
| 37 | using Base = COWHelper<IColumn, ColumnNullable>; |
| 38 | static Ptr create(const ColumnPtr & nested_column_, const ColumnPtr & null_map_) |
| 39 | { |
| 40 | return ColumnNullable::create(nested_column_->assumeMutable(), null_map_->assumeMutable()); |
| 41 | } |
| 42 | |
| 43 | template <typename ... Args, typename = typename std::enable_if<IsMutableColumns<Args ...>::value>::type> |
| 44 | static MutablePtr create(Args &&... args) { return Base::create(std::forward<Args>(args)...); } |
| 45 | |
| 46 | const char * getFamilyName() const override { return "Nullable" ; } |
| 47 | std::string getName() const override { return "Nullable(" + nested_column->getName() + ")" ; } |
| 48 | MutableColumnPtr cloneResized(size_t size) const override; |
| 49 | size_t size() const override { return nested_column->size(); } |
| 50 | bool isNullAt(size_t n) const override { return assert_cast<const ColumnUInt8 &>(*null_map).getData()[n] != 0;} |
| 51 | Field operator[](size_t n) const override; |
| 52 | void get(size_t n, Field & res) const override; |
| 53 | bool getBool(size_t n) const override { return isNullAt(n) ? 0 : nested_column->getBool(n); } |
| 54 | UInt64 get64(size_t n) const override { return nested_column->get64(n); } |
| 55 | StringRef getDataAt(size_t n) const override; |
| 56 | |
| 57 | /// Will insert null value if pos=nullptr |
| 58 | void insertData(const char * pos, size_t length) override; |
| 59 | StringRef serializeValueIntoArena(size_t n, Arena & arena, char const *& begin) const override; |
| 60 | const char * deserializeAndInsertFromArena(const char * pos) override; |
| 61 | void insertRangeFrom(const IColumn & src, size_t start, size_t length) override; |
| 62 | void insert(const Field & x) override; |
| 63 | void insertFrom(const IColumn & src, size_t n) override; |
| 64 | |
| 65 | void insertFromNotNullable(const IColumn & src, size_t n); |
| 66 | void insertRangeFromNotNullable(const IColumn & src, size_t start, size_t length); |
| 67 | void insertManyFromNotNullable(const IColumn & src, size_t position, size_t length); |
| 68 | |
| 69 | void insertDefault() override |
| 70 | { |
| 71 | getNestedColumn().insertDefault(); |
| 72 | getNullMapData().push_back(1); |
| 73 | } |
| 74 | |
| 75 | void popBack(size_t n) override; |
| 76 | ColumnPtr filter(const Filter & filt, ssize_t result_size_hint) const override; |
| 77 | ColumnPtr permute(const Permutation & perm, size_t limit) const override; |
| 78 | ColumnPtr index(const IColumn & indexes, size_t limit) const override; |
| 79 | int compareAt(size_t n, size_t m, const IColumn & rhs_, int null_direction_hint) const override; |
| 80 | void getPermutation(bool reverse, size_t limit, int null_direction_hint, Permutation & res) const override; |
| 81 | void reserve(size_t n) override; |
| 82 | size_t byteSize() const override; |
| 83 | size_t allocatedBytes() const override; |
| 84 | void protect() override; |
| 85 | ColumnPtr replicate(const Offsets & replicate_offsets) const override; |
| 86 | void updateHashWithValue(size_t n, SipHash & hash) const override; |
| 87 | void getExtremes(Field & min, Field & max) const override; |
| 88 | |
| 89 | MutableColumns scatter(ColumnIndex num_columns, const Selector & selector) const override |
| 90 | { |
| 91 | return scatterImpl<ColumnNullable>(num_columns, selector); |
| 92 | } |
| 93 | |
| 94 | void gather(ColumnGathererStream & gatherer_stream) override; |
| 95 | |
| 96 | void forEachSubcolumn(ColumnCallback callback) override |
| 97 | { |
| 98 | callback(nested_column); |
| 99 | callback(null_map); |
| 100 | } |
| 101 | |
| 102 | bool structureEquals(const IColumn & rhs) const override |
| 103 | { |
| 104 | if (auto rhs_nullable = typeid_cast<const ColumnNullable *>(&rhs)) |
| 105 | return nested_column->structureEquals(*rhs_nullable->nested_column); |
| 106 | return false; |
| 107 | } |
| 108 | |
| 109 | bool isNullable() const override { return true; } |
| 110 | bool isFixedAndContiguous() const override { return false; } |
| 111 | bool valuesHaveFixedSize() const override { return nested_column->valuesHaveFixedSize(); } |
| 112 | size_t sizeOfValueIfFixed() const override { return null_map->sizeOfValueIfFixed() + nested_column->sizeOfValueIfFixed(); } |
| 113 | bool onlyNull() const override { return nested_column->isDummy(); } |
| 114 | |
| 115 | |
| 116 | /// Return the column that represents values. |
| 117 | IColumn & getNestedColumn() { return *nested_column; } |
| 118 | const IColumn & getNestedColumn() const { return *nested_column; } |
| 119 | |
| 120 | const ColumnPtr & getNestedColumnPtr() const { return nested_column; } |
| 121 | |
| 122 | /// Return the column that represents the byte map. |
| 123 | const ColumnPtr & getNullMapColumnPtr() const { return null_map; } |
| 124 | |
| 125 | ColumnUInt8 & getNullMapColumn() { return assert_cast<ColumnUInt8 &>(*null_map); } |
| 126 | const ColumnUInt8 & getNullMapColumn() const { return assert_cast<const ColumnUInt8 &>(*null_map); } |
| 127 | |
| 128 | NullMap & getNullMapData() { return getNullMapColumn().getData(); } |
| 129 | const NullMap & getNullMapData() const { return getNullMapColumn().getData(); } |
| 130 | |
| 131 | /// Apply the null byte map of a specified nullable column onto the |
| 132 | /// null byte map of the current column by performing an element-wise OR |
| 133 | /// between both byte maps. This method is used to determine the null byte |
| 134 | /// map of the result column of a function taking one or more nullable |
| 135 | /// columns. |
| 136 | void applyNullMap(const ColumnNullable & other); |
| 137 | void applyNullMap(const ColumnUInt8 & map); |
| 138 | void applyNegatedNullMap(const ColumnUInt8 & map); |
| 139 | |
| 140 | /// Check that size of null map equals to size of nested column. |
| 141 | void checkConsistency() const; |
| 142 | |
| 143 | private: |
| 144 | WrappedPtr nested_column; |
| 145 | WrappedPtr null_map; |
| 146 | |
| 147 | template <bool negative> |
| 148 | void applyNullMapImpl(const ColumnUInt8 & map); |
| 149 | }; |
| 150 | |
| 151 | ColumnPtr makeNullable(const ColumnPtr & column); |
| 152 | |
| 153 | } |
| 154 | |