be/src/exprs/aggregate/aggregate_function_stddev.h
Line | Count | Source |
1 | | // Licensed to the Apache Software Foundation (ASF) under one |
2 | | // or more contributor license agreements. See the NOTICE file |
3 | | // distributed with this work for additional information |
4 | | // regarding copyright ownership. The ASF licenses this file |
5 | | // to you under the Apache License, Version 2.0 (the |
6 | | // "License"); you may not use this file except in compliance |
7 | | // with the License. You may obtain a copy of the License at |
8 | | // |
9 | | // http://www.apache.org/licenses/LICENSE-2.0 |
10 | | // |
11 | | // Unless required by applicable law or agreed to in writing, |
12 | | // software distributed under the License is distributed on an |
13 | | // "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY |
14 | | // KIND, either express or implied. See the License for the |
15 | | // specific language governing permissions and limitations |
16 | | // under the License. |
17 | | |
18 | | #pragma once |
19 | | |
20 | | #include <boost/iterator/iterator_facade.hpp> |
21 | | #include <cmath> |
22 | | #include <cstddef> |
23 | | #include <cstdint> |
24 | | #include <memory> |
25 | | #include <type_traits> |
26 | | |
27 | | #include "core/assert_cast.h" |
28 | | #include "core/column/column.h" |
29 | | #include "core/column/column_nullable.h" |
30 | | #include "core/data_type/data_type_decimal.h" |
31 | | #include "core/data_type/data_type_number.h" |
32 | | #include "core/types.h" |
33 | | #include "exprs/aggregate/aggregate_function.h" |
34 | | |
35 | | namespace doris { |
36 | | class Arena; |
37 | | class BufferReadable; |
38 | | class BufferWritable; |
39 | | template <PrimitiveType T> |
40 | | class ColumnDecimal; |
41 | | template <PrimitiveType T> |
42 | | class ColumnVector; |
43 | | |
44 | | template <PrimitiveType T, bool is_stddev> |
45 | | struct BaseData { |
46 | 0 | BaseData() = default; Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EEC2Ev Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EEC2Ev |
47 | 0 | virtual ~BaseData() = default; Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EED2Ev Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EED2Ev |
48 | | |
49 | 0 | void write(BufferWritable& buf) const { |
50 | 0 | buf.write_binary(mean); |
51 | 0 | buf.write_binary(m2); |
52 | 0 | buf.write_binary(count); |
53 | 0 | } Unexecuted instantiation: _ZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EE5writeERNS_14BufferWritableE Unexecuted instantiation: _ZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EE5writeERNS_14BufferWritableE |
54 | | |
55 | 0 | void read(BufferReadable& buf) { |
56 | 0 | buf.read_binary(mean); |
57 | 0 | buf.read_binary(m2); |
58 | 0 | buf.read_binary(count); |
59 | 0 | } Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EE4readERNS_14BufferReadableE Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EE4readERNS_14BufferReadableE |
60 | | |
61 | 0 | void reset() { |
62 | 0 | mean = 0.0; |
63 | 0 | m2 = 0.0; |
64 | 0 | count = 0; |
65 | 0 | } Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EE5resetEv Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EE5resetEv |
66 | | |
67 | 0 | double get_result(double res) const { |
68 | 0 | auto inf_to_nan = [](double val) { |
69 | | // This function performs squaring operations, and due to differences in computation order, |
70 | | // it might produce different values such as inf and nan. |
71 | | // In MySQL, this will directly result in an error due to exceeding the double range. |
72 | | // For performance reasons, we are uniformly changing it to nan |
73 | 0 | if (std::isinf(val)) { |
74 | 0 | return std::nan(""); |
75 | 0 | } |
76 | 0 | return val; |
77 | 0 | }; Unexecuted instantiation: _ZZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EE10get_resultEdENKUldE_clEd Unexecuted instantiation: _ZZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EE10get_resultEdENKUldE_clEd |
78 | 0 | if constexpr (is_stddev) { |
79 | 0 | return inf_to_nan(std::sqrt(res)); |
80 | 0 | } else { |
81 | 0 | return inf_to_nan(res); |
82 | 0 | } |
83 | 0 | } Unexecuted instantiation: _ZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EE10get_resultEd Unexecuted instantiation: _ZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EE10get_resultEd |
84 | | |
85 | 0 | double get_pop_result() const { |
86 | 0 | if (count == 1) { |
87 | 0 | return 0.0; |
88 | 0 | } |
89 | 0 | double res = m2 / (double)count; |
90 | 0 | return get_result(res); |
91 | 0 | } Unexecuted instantiation: _ZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EE14get_pop_resultEv Unexecuted instantiation: _ZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EE14get_pop_resultEv |
92 | | |
93 | 0 | double get_samp_result() const { |
94 | 0 | double res = m2 / double(count - 1); |
95 | 0 | return get_result(res); |
96 | 0 | } Unexecuted instantiation: _ZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EE15get_samp_resultEv Unexecuted instantiation: _ZNK5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EE15get_samp_resultEv |
97 | | |
98 | 0 | void merge(const BaseData& rhs) { |
99 | 0 | if (rhs.count == 0) { |
100 | 0 | return; |
101 | 0 | } |
102 | 0 | double delta = mean - rhs.mean; |
103 | 0 | double sum_count = double(count + rhs.count); |
104 | 0 | mean = rhs.mean + delta * (double)count / sum_count; |
105 | 0 | m2 = rhs.m2 + m2 + (delta * delta) * (double)rhs.count * (double)count / sum_count; |
106 | 0 | count = int64_t(sum_count); |
107 | 0 | } Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EE5mergeERKS2_ Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EE5mergeERKS2_ |
108 | | |
109 | 0 | void add(const IColumn* column, size_t row_num) { |
110 | 0 | const auto& sources = assert_cast<const typename PrimitiveTypeTraits<T>::ColumnType&, |
111 | 0 | TypeCheckOnRelease::DISABLE>(*column); |
112 | 0 | double source_data = (double)sources.get_data()[row_num]; |
113 | |
|
114 | 0 | double delta = source_data - mean; |
115 | 0 | double r = delta / double(1 + count); |
116 | 0 | mean += r; |
117 | 0 | m2 += (double)count * delta * r; |
118 | 0 | count += 1; |
119 | 0 | } Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb0EE3addEPKNS_7IColumnEm Unexecuted instantiation: _ZN5doris8BaseDataILNS_13PrimitiveTypeE9ELb1EE3addEPKNS_7IColumnEm |
120 | | |
121 | | double mean {}; |
122 | | double m2 {}; |
123 | | int64_t count {}; |
124 | | }; |
125 | | |
126 | | template <PrimitiveType T, typename Name, bool is_stddev> |
127 | | struct PopData : BaseData<T, is_stddev>, Name { |
128 | | using ColVecResult = std::conditional_t<is_decimal(T), ColumnDecimal128V2, ColumnFloat64>; |
129 | 0 | void insert_result_into(IColumn& to) const { |
130 | 0 | auto& col = assert_cast<ColVecResult&>(to); |
131 | | if constexpr (is_decimal(T)) { |
132 | | col.get_data().push_back(this->get_pop_result().value()); |
133 | 0 | } else { |
134 | 0 | col.get_data().push_back(this->get_pop_result()); |
135 | 0 | } |
136 | 0 | } Unexecuted instantiation: _ZNK5doris7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EE18insert_result_intoERNS_7IColumnE Unexecuted instantiation: _ZNK5doris7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EE18insert_result_intoERNS_7IColumnE |
137 | | |
138 | 0 | static DataTypePtr get_return_type() { return std::make_shared<DataTypeFloat64>(); }Unexecuted instantiation: _ZN5doris7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EE15get_return_typeEv Unexecuted instantiation: _ZN5doris7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EE15get_return_typeEv |
139 | | }; |
140 | | |
141 | | // For this series of functions, the Decimal type is not supported |
142 | | // because the operations involve squaring, |
143 | | // which can easily exceed the range of the Decimal type. |
144 | | |
145 | | template <PrimitiveType T, typename Name, bool is_stddev> |
146 | | struct SampData : BaseData<T, is_stddev>, Name { |
147 | | using ColVecResult = std::conditional_t<is_decimal(T), ColumnDecimal128V2, ColumnFloat64>; |
148 | 0 | void insert_result_into(IColumn& to) const { |
149 | 0 | auto& col = assert_cast<ColVecResult&>(to); |
150 | 0 | if (this->count == 1 || this->count == 0) { |
151 | 0 | col.insert_default(); |
152 | 0 | } else { |
153 | | if constexpr (is_decimal(T)) { |
154 | | col.get_data().push_back(this->get_samp_result().value()); |
155 | 0 | } else { |
156 | 0 | col.get_data().push_back(this->get_samp_result()); |
157 | 0 | } |
158 | 0 | } |
159 | 0 | } Unexecuted instantiation: _ZNK5doris8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EE18insert_result_intoERNS_7IColumnE Unexecuted instantiation: _ZNK5doris8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EE18insert_result_intoERNS_7IColumnE |
160 | | |
161 | 0 | static DataTypePtr get_return_type() { return std::make_shared<DataTypeFloat64>(); }Unexecuted instantiation: _ZN5doris8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EE15get_return_typeEv Unexecuted instantiation: _ZN5doris8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EE15get_return_typeEv |
162 | | }; |
163 | | |
164 | | struct StddevName { |
165 | 0 | static const char* name() { return "stddev"; } |
166 | | }; |
167 | | struct VarianceName { |
168 | 0 | static const char* name() { return "variance"; } |
169 | | }; |
170 | | struct VarianceSampName { |
171 | 0 | static const char* name() { return "variance_samp"; } |
172 | | }; |
173 | | struct StddevSampName { |
174 | 0 | static const char* name() { return "stddev_samp"; } |
175 | | }; |
176 | | |
177 | | template <typename Data> |
178 | | class AggregateFunctionSampVariance final |
179 | | : public IAggregateFunctionDataHelper<Data, AggregateFunctionSampVariance<Data>>, |
180 | | UnaryExpression, |
181 | | NullableAggregateFunction { |
182 | | public: |
183 | | AggregateFunctionSampVariance(const DataTypes& argument_types_) |
184 | 0 | : IAggregateFunctionDataHelper<Data, AggregateFunctionSampVariance<Data>>( |
185 | 0 | argument_types_) {}Unexecuted instantiation: _ZN5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EEEEC2ERKSt6vectorISt10shared_ptrIKNS_9IDataTypeEESaISA_EE Unexecuted instantiation: _ZN5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EEEEC2ERKSt6vectorISt10shared_ptrIKNS_9IDataTypeEESaISA_EE Unexecuted instantiation: _ZN5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EEEEC2ERKSt6vectorISt10shared_ptrIKNS_9IDataTypeEESaISA_EE Unexecuted instantiation: _ZN5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EEEEC2ERKSt6vectorISt10shared_ptrIKNS_9IDataTypeEESaISA_EE |
186 | | |
187 | 0 | String get_name() const override { return Data::name(); }Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EEEE8get_nameB5cxx11Ev Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EEEE8get_nameB5cxx11Ev Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EEEE8get_nameB5cxx11Ev Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EEEE8get_nameB5cxx11Ev |
188 | | |
189 | 0 | DataTypePtr get_return_type() const override { return Data::get_return_type(); }Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EEEE15get_return_typeEv Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EEEE15get_return_typeEv Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EEEE15get_return_typeEv Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EEEE15get_return_typeEv |
190 | | |
191 | | void add(AggregateDataPtr __restrict place, const IColumn** columns, ssize_t row_num, |
192 | 0 | Arena&) const override { |
193 | 0 | this->data(place).add(columns[0], row_num); |
194 | 0 | } Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EEEE3addEPcPPKNS_7IColumnElRNS_5ArenaE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EEEE3addEPcPPKNS_7IColumnElRNS_5ArenaE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EEEE3addEPcPPKNS_7IColumnElRNS_5ArenaE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EEEE3addEPcPPKNS_7IColumnElRNS_5ArenaE |
195 | | |
196 | 0 | void reset(AggregateDataPtr __restrict place) const override { this->data(place).reset(); }Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EEEE5resetEPc Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EEEE5resetEPc Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EEEE5resetEPc Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EEEE5resetEPc |
197 | | |
198 | | void merge(AggregateDataPtr __restrict place, ConstAggregateDataPtr rhs, |
199 | 0 | Arena&) const override { |
200 | 0 | this->data(place).merge(this->data(rhs)); |
201 | 0 | } Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EEEE5mergeEPcPKcRNS_5ArenaE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EEEE5mergeEPcPKcRNS_5ArenaE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EEEE5mergeEPcPKcRNS_5ArenaE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EEEE5mergeEPcPKcRNS_5ArenaE |
202 | | |
203 | 0 | void serialize(ConstAggregateDataPtr __restrict place, BufferWritable& buf) const override { |
204 | 0 | this->data(place).write(buf); |
205 | 0 | } Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EEEE9serializeEPKcRNS_14BufferWritableE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EEEE9serializeEPKcRNS_14BufferWritableE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EEEE9serializeEPKcRNS_14BufferWritableE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EEEE9serializeEPKcRNS_14BufferWritableE |
206 | | |
207 | | void deserialize(AggregateDataPtr __restrict place, BufferReadable& buf, |
208 | 0 | Arena&) const override { |
209 | 0 | this->data(place).read(buf); |
210 | 0 | } Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EEEE11deserializeEPcRNS_14BufferReadableERNS_5ArenaE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EEEE11deserializeEPcRNS_14BufferReadableERNS_5ArenaE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EEEE11deserializeEPcRNS_14BufferReadableERNS_5ArenaE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EEEE11deserializeEPcRNS_14BufferReadableERNS_5ArenaE |
211 | | |
212 | 0 | void insert_result_into(ConstAggregateDataPtr __restrict place, IColumn& to) const override { |
213 | 0 | this->data(place).insert_result_into(to); |
214 | 0 | } Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_16VarianceSampNameELb0EEEE18insert_result_intoEPKcRNS_7IColumnE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_12VarianceNameELb0EEEE18insert_result_intoEPKcRNS_7IColumnE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_7PopDataILNS_13PrimitiveTypeE9ENS_10StddevNameELb1EEEE18insert_result_intoEPKcRNS_7IColumnE Unexecuted instantiation: _ZNK5doris29AggregateFunctionSampVarianceINS_8SampDataILNS_13PrimitiveTypeE9ENS_14StddevSampNameELb1EEEE18insert_result_intoEPKcRNS_7IColumnE |
215 | | }; |
216 | | |
217 | | } // namespace doris |