be/src/exprs/function/in.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 | | // This file is copied from |
18 | | |
19 | | #pragma once |
20 | | |
21 | | #include <glog/logging.h> |
22 | | |
23 | | #include <boost/iterator/iterator_facade.hpp> |
24 | | #include <cstddef> |
25 | | #include <memory> |
26 | | #include <utility> |
27 | | #include <vector> |
28 | | |
29 | | #include "common/status.h" |
30 | | #include "core/block/block.h" |
31 | | #include "core/block/column_numbers.h" |
32 | | #include "core/block/column_with_type_and_name.h" |
33 | | #include "core/column/column.h" |
34 | | #include "core/column/column_const.h" |
35 | | #include "core/column/column_nullable.h" |
36 | | #include "core/column/column_vector.h" |
37 | | #include "core/data_type/data_type.h" |
38 | | #include "core/data_type/data_type_nullable.h" |
39 | | #include "core/data_type/data_type_number.h" |
40 | | #include "core/data_type/define_primitive_type.h" |
41 | | #include "core/field.h" |
42 | | #include "core/string_ref.h" |
43 | | #include "core/types.h" |
44 | | #include "exprs/aggregate/aggregate_function.h" |
45 | | #include "exprs/create_predicate_function.h" |
46 | | #include "exprs/function/function.h" |
47 | | #include "exprs/function_context.h" |
48 | | #include "exprs/hybrid_set.h" |
49 | | #include "storage/index/index_reader_helper.h" |
50 | | |
51 | | namespace doris { |
52 | | |
53 | | template <typename T> |
54 | | class ColumnStr; |
55 | | using ColumnString = ColumnStr<UInt32>; |
56 | | |
57 | | struct InState { |
58 | | bool use_set = true; |
59 | | std::shared_ptr<HybridSetBase> hybrid_set; |
60 | | }; |
61 | | |
62 | | template <bool negative> |
63 | | class FunctionIn : public IFunction { |
64 | | public: |
65 | | static constexpr auto name = negative ? "not_in" : "in"; |
66 | | |
67 | 20 | static FunctionPtr create() { return std::make_shared<FunctionIn>(); }_ZN5doris10FunctionInILb0EE6createEv Line | Count | Source | 67 | 12 | static FunctionPtr create() { return std::make_shared<FunctionIn>(); } |
_ZN5doris10FunctionInILb1EE6createEv Line | Count | Source | 67 | 8 | static FunctionPtr create() { return std::make_shared<FunctionIn>(); } |
|
68 | | |
69 | 0 | String get_name() const override { return name; }Unexecuted instantiation: _ZNK5doris10FunctionInILb0EE8get_nameB5cxx11Ev Unexecuted instantiation: _ZNK5doris10FunctionInILb1EE8get_nameB5cxx11Ev |
70 | | |
71 | 18 | bool is_variadic() const override { return true; }_ZNK5doris10FunctionInILb0EE11is_variadicEv Line | Count | Source | 71 | 11 | bool is_variadic() const override { return true; } |
_ZNK5doris10FunctionInILb1EE11is_variadicEv Line | Count | Source | 71 | 7 | bool is_variadic() const override { return true; } |
|
72 | | |
73 | 0 | size_t get_number_of_arguments() const override { return 0; }Unexecuted instantiation: _ZNK5doris10FunctionInILb0EE23get_number_of_argumentsEv Unexecuted instantiation: _ZNK5doris10FunctionInILb1EE23get_number_of_argumentsEv |
74 | | |
75 | 16 | DataTypePtr get_return_type_impl(const DataTypes& args) const override { |
76 | 22 | for (const auto& arg : args) { |
77 | 22 | if (arg->is_nullable()) { |
78 | 16 | return make_nullable(std::make_shared<DataTypeUInt8>()); |
79 | 16 | } |
80 | 22 | } |
81 | 0 | return std::make_shared<DataTypeUInt8>(); |
82 | 16 | } _ZNK5doris10FunctionInILb0EE20get_return_type_implERKSt6vectorISt10shared_ptrIKNS_9IDataTypeEESaIS6_EE Line | Count | Source | 75 | 10 | DataTypePtr get_return_type_impl(const DataTypes& args) const override { | 76 | 14 | for (const auto& arg : args) { | 77 | 14 | if (arg->is_nullable()) { | 78 | 10 | return make_nullable(std::make_shared<DataTypeUInt8>()); | 79 | 10 | } | 80 | 14 | } | 81 | 0 | return std::make_shared<DataTypeUInt8>(); | 82 | 10 | } |
_ZNK5doris10FunctionInILb1EE20get_return_type_implERKSt6vectorISt10shared_ptrIKNS_9IDataTypeEESaIS6_EE Line | Count | Source | 75 | 6 | DataTypePtr get_return_type_impl(const DataTypes& args) const override { | 76 | 8 | for (const auto& arg : args) { | 77 | 8 | if (arg->is_nullable()) { | 78 | 6 | return make_nullable(std::make_shared<DataTypeUInt8>()); | 79 | 6 | } | 80 | 8 | } | 81 | 0 | return std::make_shared<DataTypeUInt8>(); | 82 | 6 | } |
|
83 | | |
84 | 16 | bool use_default_implementation_for_nulls() const override { return false; }_ZNK5doris10FunctionInILb0EE36use_default_implementation_for_nullsEv Line | Count | Source | 84 | 10 | bool use_default_implementation_for_nulls() const override { return false; } |
_ZNK5doris10FunctionInILb1EE36use_default_implementation_for_nullsEv Line | Count | Source | 84 | 6 | bool use_default_implementation_for_nulls() const override { return false; } |
|
85 | | |
86 | | // size of [ in ( 1 , 2 , 3 , null) ] is 3 |
87 | 11 | size_t get_size_with_out_null(FunctionContext* context) { |
88 | 11 | if ((context->get_num_args() - 1) > FIXED_CONTAINER_MAX_SIZE) { |
89 | 0 | return context->get_num_args() - 1; |
90 | 0 | } |
91 | 11 | size_t sz = 0; |
92 | 34 | for (int i = 1; i < context->get_num_args(); ++i) { |
93 | 23 | const auto& const_column_ptr = context->get_constant_col(i); |
94 | 23 | if (const_column_ptr != nullptr) { |
95 | 23 | auto const_data = const_column_ptr->column_ptr->get_data_at(0); |
96 | 23 | if (const_data.data != nullptr) { |
97 | 21 | sz++; |
98 | 21 | } |
99 | 23 | } |
100 | 23 | } |
101 | 11 | return sz; |
102 | 11 | } _ZN5doris10FunctionInILb0EE22get_size_with_out_nullEPNS_15FunctionContextE Line | Count | Source | 87 | 7 | size_t get_size_with_out_null(FunctionContext* context) { | 88 | 7 | if ((context->get_num_args() - 1) > FIXED_CONTAINER_MAX_SIZE) { | 89 | 0 | return context->get_num_args() - 1; | 90 | 0 | } | 91 | 7 | size_t sz = 0; | 92 | 24 | for (int i = 1; i < context->get_num_args(); ++i) { | 93 | 17 | const auto& const_column_ptr = context->get_constant_col(i); | 94 | 17 | if (const_column_ptr != nullptr) { | 95 | 17 | auto const_data = const_column_ptr->column_ptr->get_data_at(0); | 96 | 17 | if (const_data.data != nullptr) { | 97 | 16 | sz++; | 98 | 16 | } | 99 | 17 | } | 100 | 17 | } | 101 | 7 | return sz; | 102 | 7 | } |
_ZN5doris10FunctionInILb1EE22get_size_with_out_nullEPNS_15FunctionContextE Line | Count | Source | 87 | 4 | size_t get_size_with_out_null(FunctionContext* context) { | 88 | 4 | if ((context->get_num_args() - 1) > FIXED_CONTAINER_MAX_SIZE) { | 89 | 0 | return context->get_num_args() - 1; | 90 | 0 | } | 91 | 4 | size_t sz = 0; | 92 | 10 | for (int i = 1; i < context->get_num_args(); ++i) { | 93 | 6 | const auto& const_column_ptr = context->get_constant_col(i); | 94 | 6 | if (const_column_ptr != nullptr) { | 95 | 6 | auto const_data = const_column_ptr->column_ptr->get_data_at(0); | 96 | 6 | if (const_data.data != nullptr) { | 97 | 5 | sz++; | 98 | 5 | } | 99 | 6 | } | 100 | 6 | } | 101 | 4 | return sz; | 102 | 4 | } |
|
103 | | |
104 | 26 | Status open(FunctionContext* context, FunctionContext::FunctionStateScope scope) override { |
105 | 26 | if (scope == FunctionContext::THREAD_LOCAL) { |
106 | 11 | return Status::OK(); |
107 | 11 | } |
108 | | // Binary IO must not route IN through the shared string/storage predicate implementation. |
109 | 15 | if (context->get_arg_type(0)->get_primitive_type() == TYPE_VARBINARY) { |
110 | 4 | return Status::NotSupported("VARBINARY IN/NOT IN is not supported"); |
111 | 4 | } |
112 | 11 | std::shared_ptr<InState> state = std::make_shared<InState>(); |
113 | 11 | context->set_function_state(scope, state); |
114 | 11 | DCHECK(context->get_num_args() >= 1); |
115 | 11 | if (context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_NULL) { |
116 | 0 | state->hybrid_set.reset(create_set(TYPE_BOOLEAN, 0, true)); |
117 | 11 | } else if (context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_CHAR || |
118 | 11 | context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_VARCHAR || |
119 | 11 | context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_STRING) { |
120 | | // the StringValue's memory is held by FunctionContext, so we can use StringValueSet here directly |
121 | 3 | state->hybrid_set.reset(create_string_value_set(get_size_with_out_null(context), true)); |
122 | 8 | } else { |
123 | 8 | state->hybrid_set.reset(create_set(context->get_arg_type(0)->get_primitive_type(), |
124 | 8 | get_size_with_out_null(context), true)); |
125 | 8 | } |
126 | | |
127 | 34 | for (int i = 1; i < context->get_num_args(); ++i) { |
128 | 23 | const auto& const_column_ptr = context->get_constant_col(i); |
129 | 23 | if (const_column_ptr != nullptr) { |
130 | 23 | auto const_data = const_column_ptr->column_ptr->get_data_at(0); |
131 | 23 | state->hybrid_set->insert((void*)const_data.data, const_data.size); |
132 | 23 | } else { |
133 | 0 | state->use_set = false; |
134 | 0 | state->hybrid_set.reset(); |
135 | 0 | break; |
136 | 0 | } |
137 | 23 | } |
138 | 11 | return Status::OK(); |
139 | 15 | } _ZN5doris10FunctionInILb0EE4openEPNS_15FunctionContextENS2_18FunctionStateScopeE Line | Count | Source | 104 | 16 | Status open(FunctionContext* context, FunctionContext::FunctionStateScope scope) override { | 105 | 16 | if (scope == FunctionContext::THREAD_LOCAL) { | 106 | 7 | return Status::OK(); | 107 | 7 | } | 108 | | // Binary IO must not route IN through the shared string/storage predicate implementation. | 109 | 9 | if (context->get_arg_type(0)->get_primitive_type() == TYPE_VARBINARY) { | 110 | 2 | return Status::NotSupported("VARBINARY IN/NOT IN is not supported"); | 111 | 2 | } | 112 | 7 | std::shared_ptr<InState> state = std::make_shared<InState>(); | 113 | 7 | context->set_function_state(scope, state); | 114 | 7 | DCHECK(context->get_num_args() >= 1); | 115 | 7 | if (context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_NULL) { | 116 | 0 | state->hybrid_set.reset(create_set(TYPE_BOOLEAN, 0, true)); | 117 | 7 | } else if (context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_CHAR || | 118 | 7 | context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_VARCHAR || | 119 | 7 | context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_STRING) { | 120 | | // the StringValue's memory is held by FunctionContext, so we can use StringValueSet here directly | 121 | 2 | state->hybrid_set.reset(create_string_value_set(get_size_with_out_null(context), true)); | 122 | 5 | } else { | 123 | 5 | state->hybrid_set.reset(create_set(context->get_arg_type(0)->get_primitive_type(), | 124 | 5 | get_size_with_out_null(context), true)); | 125 | 5 | } | 126 | | | 127 | 24 | for (int i = 1; i < context->get_num_args(); ++i) { | 128 | 17 | const auto& const_column_ptr = context->get_constant_col(i); | 129 | 17 | if (const_column_ptr != nullptr) { | 130 | 17 | auto const_data = const_column_ptr->column_ptr->get_data_at(0); | 131 | 17 | state->hybrid_set->insert((void*)const_data.data, const_data.size); | 132 | 17 | } else { | 133 | 0 | state->use_set = false; | 134 | 0 | state->hybrid_set.reset(); | 135 | 0 | break; | 136 | 0 | } | 137 | 17 | } | 138 | 7 | return Status::OK(); | 139 | 9 | } |
_ZN5doris10FunctionInILb1EE4openEPNS_15FunctionContextENS2_18FunctionStateScopeE Line | Count | Source | 104 | 10 | Status open(FunctionContext* context, FunctionContext::FunctionStateScope scope) override { | 105 | 10 | if (scope == FunctionContext::THREAD_LOCAL) { | 106 | 4 | return Status::OK(); | 107 | 4 | } | 108 | | // Binary IO must not route IN through the shared string/storage predicate implementation. | 109 | 6 | if (context->get_arg_type(0)->get_primitive_type() == TYPE_VARBINARY) { | 110 | 2 | return Status::NotSupported("VARBINARY IN/NOT IN is not supported"); | 111 | 2 | } | 112 | 4 | std::shared_ptr<InState> state = std::make_shared<InState>(); | 113 | 4 | context->set_function_state(scope, state); | 114 | 4 | DCHECK(context->get_num_args() >= 1); | 115 | 4 | if (context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_NULL) { | 116 | 0 | state->hybrid_set.reset(create_set(TYPE_BOOLEAN, 0, true)); | 117 | 4 | } else if (context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_CHAR || | 118 | 4 | context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_VARCHAR || | 119 | 4 | context->get_arg_type(0)->get_primitive_type() == PrimitiveType::TYPE_STRING) { | 120 | | // the StringValue's memory is held by FunctionContext, so we can use StringValueSet here directly | 121 | 1 | state->hybrid_set.reset(create_string_value_set(get_size_with_out_null(context), true)); | 122 | 3 | } else { | 123 | 3 | state->hybrid_set.reset(create_set(context->get_arg_type(0)->get_primitive_type(), | 124 | 3 | get_size_with_out_null(context), true)); | 125 | 3 | } | 126 | | | 127 | 10 | for (int i = 1; i < context->get_num_args(); ++i) { | 128 | 6 | const auto& const_column_ptr = context->get_constant_col(i); | 129 | 6 | if (const_column_ptr != nullptr) { | 130 | 6 | auto const_data = const_column_ptr->column_ptr->get_data_at(0); | 131 | 6 | state->hybrid_set->insert((void*)const_data.data, const_data.size); | 132 | 6 | } else { | 133 | 0 | state->use_set = false; | 134 | 0 | state->hybrid_set.reset(); | 135 | 0 | break; | 136 | 0 | } | 137 | 6 | } | 138 | 4 | return Status::OK(); | 139 | 6 | } |
|
140 | | |
141 | | Status evaluate_inverted_index( |
142 | | const ColumnsWithTypeAndName& arguments, |
143 | | const std::vector<IndexFieldNameAndTypePair>& data_type_with_names, |
144 | | std::vector<segment_v2::IndexIterator*> iterators, uint32_t num_rows, |
145 | | const InvertedIndexAnalyzerCtx* analyzer_ctx, |
146 | 0 | segment_v2::InvertedIndexResultBitmap& bitmap_result) const override { |
147 | 0 | DCHECK(data_type_with_names.size() == 1); |
148 | 0 | DCHECK(iterators.size() == 1); |
149 | 0 | auto* iter = iterators[0]; |
150 | 0 | auto data_type_with_name = data_type_with_names[0]; |
151 | 0 | std::shared_ptr<roaring::Roaring> roaring = std::make_shared<roaring::Roaring>(); |
152 | 0 | std::shared_ptr<roaring::Roaring> null_bitmap = std::make_shared<roaring::Roaring>(); |
153 | |
|
154 | 0 | if (iter == nullptr) { |
155 | 0 | return Status::OK(); |
156 | 0 | } |
157 | 0 | if (!segment_v2::IndexReaderHelper::has_string_or_bkd_index(iter)) { |
158 | | //NOT support in list when parser is FULLTEXT for expr inverted index evaluate. |
159 | 0 | return Status::OK(); |
160 | 0 | } |
161 | 0 | if (iter->has_null()) { |
162 | 0 | segment_v2::InvertedIndexQueryCacheHandle null_bitmap_cache_handle; |
163 | 0 | RETURN_IF_ERROR(iter->read_null_bitmap(&null_bitmap_cache_handle)); |
164 | 0 | null_bitmap = null_bitmap_cache_handle.get_bitmap(); |
165 | 0 | } |
166 | 0 | for (const auto& arg : arguments) { |
167 | 0 | Field param_value; |
168 | 0 | arg.column->get(0, param_value); |
169 | 0 | if (param_value.is_null()) { |
170 | | // predicate like column NOT IN (NULL, '') should not push down to index. |
171 | 0 | if (negative) { |
172 | 0 | return Status::OK(); |
173 | 0 | } |
174 | 0 | *roaring |= *null_bitmap; |
175 | 0 | continue; |
176 | 0 | } |
177 | 0 | InvertedIndexQueryType query_type = InvertedIndexQueryType::EQUAL_QUERY; |
178 | 0 | segment_v2::InvertedIndexParam param; |
179 | 0 | param.column_name = data_type_with_name.first; |
180 | 0 | param.column_type = data_type_with_name.second; |
181 | 0 | param.query_value = param_value; |
182 | 0 | param.query_type = query_type; |
183 | 0 | param.num_rows = num_rows; |
184 | 0 | param.roaring = std::make_shared<roaring::Roaring>(); |
185 | 0 | param.analyzer_ctx = analyzer_ctx; |
186 | 0 | RETURN_IF_ERROR(iter->read_from_index(segment_v2::IndexParam {¶m})); |
187 | 0 | *roaring |= *param.roaring; |
188 | 0 | } |
189 | 0 | segment_v2::InvertedIndexResultBitmap result(roaring, null_bitmap); |
190 | 0 | bitmap_result = result; |
191 | 0 | bitmap_result.mask_out_null(); |
192 | 0 | if constexpr (negative) { |
193 | 0 | roaring::Roaring full_result; |
194 | 0 | full_result.addRange(0, num_rows); |
195 | 0 | bitmap_result.op_not(&full_result); |
196 | 0 | } |
197 | 0 | return Status::OK(); |
198 | 0 | } Unexecuted instantiation: _ZNK5doris10FunctionInILb0EE23evaluate_inverted_indexERKSt6vectorINS_21ColumnWithTypeAndNameESaIS3_EERKS2_ISt4pairINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEESt10shared_ptrIKNS_9IDataTypeEEESaISJ_EES2_IPNS_10segment_v213IndexIteratorESaISQ_EEjPKNS_24InvertedIndexAnalyzerCtxERNSO_25InvertedIndexResultBitmapE Unexecuted instantiation: _ZNK5doris10FunctionInILb1EE23evaluate_inverted_indexERKSt6vectorINS_21ColumnWithTypeAndNameESaIS3_EERKS2_ISt4pairINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEESt10shared_ptrIKNS_9IDataTypeEEESaISJ_EES2_IPNS_10segment_v213IndexIteratorESaISQ_EEjPKNS_24InvertedIndexAnalyzerCtxERNSO_25InvertedIndexResultBitmapE |
199 | | |
200 | | Status execute_impl(FunctionContext* context, Block& block, const ColumnNumbers& arguments, |
201 | 0 | uint32_t result, size_t input_rows_count) const override { |
202 | 0 | auto* in_state = reinterpret_cast<InState*>( |
203 | 0 | context->get_function_state(FunctionContext::FRAGMENT_LOCAL)); |
204 | 0 | if (!in_state) { |
205 | 0 | return Status::RuntimeError("funciton context for function '{}' must have Set;", |
206 | 0 | get_name()); |
207 | 0 | } |
208 | 0 | auto res = ColumnUInt8::create(); |
209 | 0 | ColumnUInt8::Container& vec_res = res->get_data(); |
210 | 0 | vec_res.resize(input_rows_count); |
211 | |
|
212 | 0 | ColumnUInt8::MutablePtr col_null_map_to; |
213 | 0 | col_null_map_to = ColumnUInt8::create(input_rows_count, false); |
214 | 0 | auto& vec_null_map_to = col_null_map_to->get_data(); |
215 | |
|
216 | 0 | const ColumnWithTypeAndName& left_arg = block.get_by_position(arguments[0]); |
217 | 0 | const auto& [unpacked_column, col_const] = unpack_if_const(left_arg.column); |
218 | 0 | ColumnPtr materialized_column = unpacked_column; |
219 | 0 | if (in_state->use_set && col_const) { |
220 | 0 | materialized_column = left_arg.column->convert_to_full_column_if_const(); |
221 | 0 | } |
222 | |
|
223 | 0 | if (in_state->use_set) { |
224 | 0 | if (const auto* null_col_ptr = |
225 | 0 | check_and_get_column<ColumnNullable>(materialized_column.get())) { |
226 | 0 | const auto& null_map = null_col_ptr->get_null_map_column().get_data(); |
227 | 0 | const auto* nested_col_ptr = null_col_ptr->get_nested_column_ptr().get(); |
228 | |
|
229 | 0 | if (nested_col_ptr->is_column_string()) { |
230 | 0 | const auto* column_string_ptr = |
231 | 0 | assert_cast<const ColumnString*>(nested_col_ptr); |
232 | 0 | search_hash_set_check_null(in_state, input_rows_count, vec_res, null_map, |
233 | 0 | column_string_ptr); |
234 | 0 | } else { |
235 | | //TODO: support other column type |
236 | 0 | search_hash_set_check_null(in_state, input_rows_count, vec_res, null_map, |
237 | 0 | nested_col_ptr); |
238 | 0 | } |
239 | |
|
240 | 0 | if (!in_state->hybrid_set->contain_null()) { |
241 | 0 | for (size_t i = 0; i < input_rows_count; ++i) { |
242 | 0 | vec_null_map_to[i] = null_map[i]; |
243 | 0 | } |
244 | 0 | } else { |
245 | 0 | for (size_t i = 0; i < input_rows_count; ++i) { |
246 | 0 | vec_null_map_to[i] = null_map[i] || negative == vec_res[i]; |
247 | 0 | } |
248 | 0 | } |
249 | |
|
250 | 0 | } else { // non-nullable |
251 | 0 | if (is_string_type(left_arg.type->get_primitive_type())) { |
252 | 0 | const auto* column_string_ptr = |
253 | 0 | assert_cast<const ColumnString*>(materialized_column.get()); |
254 | 0 | search_hash_set(in_state, input_rows_count, vec_res, column_string_ptr); |
255 | 0 | } else { |
256 | 0 | search_hash_set(in_state, input_rows_count, vec_res, materialized_column.get()); |
257 | 0 | } |
258 | |
|
259 | 0 | if (in_state->hybrid_set->contain_null()) { |
260 | 0 | for (size_t i = 0; i < input_rows_count; ++i) { |
261 | 0 | vec_null_map_to[i] = negative == vec_res[i]; |
262 | 0 | } |
263 | 0 | } |
264 | 0 | } |
265 | 0 | } else { //!in_state->use_set |
266 | 0 | std::vector<ColumnPtr> set_columns; |
267 | 0 | for (int i = 1; i < arguments.size(); ++i) { |
268 | 0 | set_columns.emplace_back(block.get_by_position(arguments[i]).column); |
269 | 0 | } |
270 | 0 | if (col_const) { |
271 | 0 | impl_without_set<true>(context, set_columns, input_rows_count, vec_res, |
272 | 0 | vec_null_map_to, materialized_column); |
273 | 0 | } else { |
274 | 0 | impl_without_set<false>(context, set_columns, input_rows_count, vec_res, |
275 | 0 | vec_null_map_to, materialized_column); |
276 | 0 | } |
277 | 0 | } |
278 | |
|
279 | 0 | if (block.get_by_position(result).type->is_nullable()) { |
280 | 0 | block.replace_by_position( |
281 | 0 | result, ColumnNullable::create(std::move(res), std::move(col_null_map_to))); |
282 | 0 | } else { |
283 | 0 | block.replace_by_position(result, std::move(res)); |
284 | 0 | } |
285 | |
|
286 | 0 | return Status::OK(); |
287 | 0 | } Unexecuted instantiation: _ZNK5doris10FunctionInILb0EE12execute_implEPNS_15FunctionContextERNS_5BlockERKSt6vectorIjSaIjEEjm Unexecuted instantiation: _ZNK5doris10FunctionInILb1EE12execute_implEPNS_15FunctionContextERNS_5BlockERKSt6vectorIjSaIjEEjm |
288 | | |
289 | | private: |
290 | | template <typename T> |
291 | | static void search_hash_set_check_null(InState* in_state, size_t input_rows_count, |
292 | | ColumnUInt8::Container& vec_res, |
293 | 0 | const ColumnUInt8::Container& null_map, T* col_ptr) { |
294 | 0 | if constexpr (!negative) { |
295 | 0 | in_state->hybrid_set->find_batch_nullable(*col_ptr, input_rows_count, null_map, |
296 | 0 | vec_res); |
297 | 0 | } else { |
298 | 0 | in_state->hybrid_set->find_batch_nullable_negative(*col_ptr, input_rows_count, null_map, |
299 | 0 | vec_res); |
300 | 0 | } |
301 | 0 | } Unexecuted instantiation: _ZN5doris10FunctionInILb0EE26search_hash_set_check_nullIKNS_9ColumnStrIjEEEEvPNS_7InStateEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEERKSC_PT_ Unexecuted instantiation: _ZN5doris10FunctionInILb0EE26search_hash_set_check_nullIKNS_7IColumnEEEvPNS_7InStateEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEERKSB_PT_ Unexecuted instantiation: _ZN5doris10FunctionInILb1EE26search_hash_set_check_nullIKNS_9ColumnStrIjEEEEvPNS_7InStateEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEERKSC_PT_ Unexecuted instantiation: _ZN5doris10FunctionInILb1EE26search_hash_set_check_nullIKNS_7IColumnEEEvPNS_7InStateEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEERKSB_PT_ |
302 | | |
303 | | template <typename T> |
304 | | static void search_hash_set(InState* in_state, size_t input_rows_count, |
305 | 0 | ColumnUInt8::Container& vec_res, T* col_ptr) { |
306 | 0 | if constexpr (!negative) { |
307 | 0 | in_state->hybrid_set->find_batch(*col_ptr, input_rows_count, vec_res); |
308 | 0 | } else { |
309 | 0 | in_state->hybrid_set->find_batch_negative(*col_ptr, input_rows_count, vec_res); |
310 | 0 | } |
311 | 0 | } Unexecuted instantiation: _ZN5doris10FunctionInILb0EE15search_hash_setIKNS_9ColumnStrIjEEEEvPNS_7InStateEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEEPT_ Unexecuted instantiation: _ZN5doris10FunctionInILb0EE15search_hash_setIKNS_7IColumnEEEvPNS_7InStateEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEEPT_ Unexecuted instantiation: _ZN5doris10FunctionInILb1EE15search_hash_setIKNS_9ColumnStrIjEEEEvPNS_7InStateEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEEPT_ Unexecuted instantiation: _ZN5doris10FunctionInILb1EE15search_hash_setIKNS_7IColumnEEEvPNS_7InStateEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEEPT_ |
312 | | |
313 | | template <bool Const> |
314 | | static void impl_without_set(FunctionContext* context, |
315 | | const std::vector<ColumnPtr>& set_columns, size_t input_rows_count, |
316 | | ColumnUInt8::Container& vec_res, |
317 | | ColumnUInt8::Container& vec_null_map_to, |
318 | 0 | const ColumnPtr& materialized_column) { |
319 | 0 | for (size_t i = 0; i < input_rows_count; ++i) { |
320 | 0 | const auto& ref_data = materialized_column->get_data_at(index_check_const(i, Const)); |
321 | 0 | if (ref_data.data == nullptr) { |
322 | 0 | vec_null_map_to[i] = true; |
323 | 0 | continue; |
324 | 0 | } |
325 | | |
326 | 0 | std::vector<StringRef> set_datas; |
327 | | // To comply with the SQL standard, IN() returns NULL not only if the expression on the left hand side is NULL, |
328 | | // but also if no match is found in the list and one of the expressions in the list is NULL. |
329 | 0 | bool null_in_set = false; |
330 | |
|
331 | 0 | for (const auto& set_column : set_columns) { |
332 | 0 | auto set_data = set_column->get_data_at(i); |
333 | 0 | if (set_data.data == nullptr) { |
334 | 0 | null_in_set = true; |
335 | 0 | } else { |
336 | 0 | set_datas.push_back(set_data); |
337 | 0 | } |
338 | 0 | } |
339 | 0 | std::unique_ptr<HybridSetBase> hybrid_set(create_set( |
340 | 0 | context->get_arg_type(0)->get_primitive_type(), set_datas.size(), true)); |
341 | 0 | for (auto& set_data : set_datas) { |
342 | 0 | hybrid_set->insert((void*)(set_data.data), set_data.size); |
343 | 0 | } |
344 | |
|
345 | 0 | vec_res[i] = negative ^ hybrid_set->find((void*)ref_data.data, ref_data.size); |
346 | 0 | if (null_in_set) { |
347 | 0 | vec_null_map_to[i] = negative == vec_res[i]; |
348 | 0 | } else { |
349 | 0 | vec_null_map_to[i] = false; |
350 | 0 | } |
351 | 0 | } |
352 | 0 | } Unexecuted instantiation: _ZN5doris10FunctionInILb0EE16impl_without_setILb1EEEvPNS_15FunctionContextERKSt6vectorINS_3COWINS_7IColumnEE13immutable_ptrIS7_EESaISA_EEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEESK_RKSA_ Unexecuted instantiation: _ZN5doris10FunctionInILb0EE16impl_without_setILb0EEEvPNS_15FunctionContextERKSt6vectorINS_3COWINS_7IColumnEE13immutable_ptrIS7_EESaISA_EEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEESK_RKSA_ Unexecuted instantiation: _ZN5doris10FunctionInILb1EE16impl_without_setILb1EEEvPNS_15FunctionContextERKSt6vectorINS_3COWINS_7IColumnEE13immutable_ptrIS7_EESaISA_EEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEESK_RKSA_ Unexecuted instantiation: _ZN5doris10FunctionInILb1EE16impl_without_setILb0EEEvPNS_15FunctionContextERKSt6vectorINS_3COWINS_7IColumnEE13immutable_ptrIS7_EESaISA_EEmRNS_8PODArrayIhLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEESK_RKSA_ |
353 | | }; |
354 | | |
355 | | } // namespace doris |