/root/doris/be/src/exprs/vtopn_pred.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 <gen_cpp/types.pb.h> |
21 | | |
22 | | #include <utility> |
23 | | |
24 | | #include "core/block/column_numbers.h" |
25 | | #include "core/data_type/data_type.h" |
26 | | #include "exec/common/util.hpp" |
27 | | #include "exprs/function/simple_function_factory.h" |
28 | | #include "exprs/vectorized_fn_call.h" |
29 | | #include "exprs/vexpr.h" |
30 | | #include "exprs/vslot_ref.h" |
31 | | #include "runtime/query_context.h" |
32 | | #include "runtime/runtime_predicate.h" |
33 | | #include "runtime/runtime_state.h" |
34 | | |
35 | | namespace doris { |
36 | | #include "common/compile_check_begin.h" |
37 | | |
38 | | // only used for dynamic topn filter |
39 | | class VTopNPred : public VExpr { |
40 | | ENABLE_FACTORY_CREATOR(VTopNPred); |
41 | | |
42 | | public: |
43 | | VTopNPred(const TExprNode& node, int source_node_id, VExprContextSPtr target_ctx) |
44 | 0 | : VExpr(node), |
45 | 0 | _source_node_id(source_node_id), |
46 | 0 | _expr_name(fmt::format("VTopNPred(source_node_id={})", _source_node_id)), |
47 | 0 | _target_ctx(std::move(target_ctx)) {} |
48 | 0 | bool is_topn_filter() const override { return true; } |
49 | | |
50 | 0 | static Status create_vtopn_pred(const TExpr& target_expr, int source_node_id, VExprSPtr& expr) { |
51 | 0 | VExprContextSPtr target_ctx; |
52 | 0 | RETURN_IF_ERROR(VExpr::create_expr_tree(target_expr, target_ctx)); |
53 | | |
54 | 0 | TExprNode node; |
55 | 0 | node.__set_node_type(TExprNodeType::FUNCTION_CALL); |
56 | 0 | node.__set_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN)); |
57 | 0 | node.__set_is_nullable(target_ctx->root()->is_nullable()); |
58 | 0 | expr = VTopNPred::create_shared(node, source_node_id, target_ctx); |
59 | |
|
60 | 0 | DCHECK(target_ctx->root() != nullptr); |
61 | 0 | expr->add_child(target_ctx->root()); |
62 | |
|
63 | 0 | return Status::OK(); |
64 | 0 | } |
65 | | |
66 | 0 | int source_node_id() const { return _source_node_id; } |
67 | 0 | Status clone_node(VExprSPtr* cloned_expr) const override { |
68 | 0 | DORIS_CHECK(cloned_expr != nullptr); |
69 | 0 | *cloned_expr = VTopNPred::create_shared(clone_texpr_node(), _source_node_id, nullptr); |
70 | 0 | return Status::OK(); |
71 | 0 | } |
72 | | |
73 | 0 | Status prepare(RuntimeState* state, const RowDescriptor& desc, VExprContext* context) override { |
74 | 0 | _predicate = &state->get_query_ctx()->get_runtime_predicate(_source_node_id); |
75 | 0 | RETURN_IF_ERROR_OR_PREPARED(VExpr::prepare(state, desc, context)); |
76 | |
|
77 | 0 | ColumnsWithTypeAndName argument_template; |
78 | 0 | argument_template.emplace_back(nullptr, _children[0]->data_type(), |
79 | 0 | _children[0]->expr_name()); |
80 | 0 | argument_template.emplace_back(nullptr, _children[0]->data_type(), "topn value"); |
81 | |
|
82 | 0 | _function = SimpleFunctionFactory::instance().get_function( |
83 | 0 | _predicate->is_asc() ? "le" : "ge", argument_template, _data_type, {}, |
84 | 0 | state->be_exec_version()); |
85 | 0 | if (!_function) { |
86 | 0 | return Status::InternalError("get function failed"); |
87 | 0 | } |
88 | 0 | return Status::OK(); |
89 | 0 | } |
90 | | |
91 | | Status execute_column(VExprContext* context, const Block* block, Selector* selector, |
92 | 0 | size_t count, ColumnPtr& result_column) const override { |
93 | 0 | if (!_predicate->has_value()) { |
94 | 0 | result_column = create_always_true_column(count, _data_type->is_nullable()); |
95 | 0 | return Status::OK(); |
96 | 0 | } |
97 | | |
98 | 0 | Block temp_block; |
99 | | |
100 | | // slot |
101 | 0 | ColumnPtr slot_column; |
102 | 0 | RETURN_IF_ERROR(_children[0]->execute_column(context, block, selector, count, slot_column)); |
103 | 0 | auto slot_type = _children[0]->execute_type(block); |
104 | 0 | temp_block.insert({slot_column, slot_type, _children[0]->expr_name()}); |
105 | 0 | int slot_id = 0; |
106 | | |
107 | | // topn value |
108 | 0 | Field field = _predicate->get_value(); |
109 | 0 | auto column_ptr = _children[0]->data_type()->create_column_const(1, field); |
110 | 0 | int topn_value_id = VExpr::insert_param(&temp_block, |
111 | 0 | {column_ptr, _children[0]->data_type(), _expr_name}, |
112 | 0 | std::max(count, column_ptr->size())); |
113 | | |
114 | | // if error(slot_id == -1), will return. |
115 | 0 | ColumnNumbers arguments = {static_cast<uint32_t>(slot_id), |
116 | 0 | static_cast<uint32_t>(topn_value_id)}; |
117 | |
|
118 | 0 | uint32_t num_columns_without_result = temp_block.columns(); |
119 | | // prepare a column to save result |
120 | 0 | temp_block.insert({nullptr, _data_type, _expr_name}); |
121 | |
|
122 | 0 | RETURN_IF_ERROR(_function->execute(nullptr, temp_block, arguments, |
123 | 0 | num_columns_without_result, temp_block.rows())); |
124 | 0 | result_column = std::move(temp_block.get_by_position(num_columns_without_result).column); |
125 | 0 | if (is_nullable() && _predicate->nulls_first()) { |
126 | | // null values are always not filtered |
127 | 0 | change_null_to_true(result_column); |
128 | 0 | } |
129 | 0 | DCHECK_EQ(result_column->size(), count); |
130 | 0 | return Status::OK(); |
131 | 0 | } |
132 | | |
133 | 0 | const std::string& expr_name() const override { return _expr_name; } |
134 | | |
135 | | // only used in external table (for min-max filter). get `slot > xxx`, not `function(slot) > xxx`. |
136 | 0 | bool get_binary_expr(VExprSPtr& new_root) const { |
137 | 0 | if (!get_child(0)->is_slot_ref()) { |
138 | | // top rf maybe is `xxx order by abs(column) limit xxx`. |
139 | 0 | return false; |
140 | 0 | } |
141 | | |
142 | 0 | if (!_predicate->has_value()) { |
143 | 0 | return false; |
144 | 0 | } |
145 | | |
146 | 0 | auto* slot_ref = assert_cast<VSlotRef*>(get_child(0).get()); |
147 | 0 | auto slot_data_type = remove_nullable(slot_ref->data_type()); |
148 | 0 | { |
149 | 0 | TFunction fn; |
150 | 0 | TFunctionName fn_name; |
151 | 0 | fn_name.__set_db_name(""); |
152 | 0 | fn_name.__set_function_name(_predicate->is_asc() ? "le" : "ge"); |
153 | 0 | fn.__set_name(fn_name); |
154 | 0 | fn.__set_binary_type(TFunctionBinaryType::BUILTIN); |
155 | 0 | std::vector<TTypeDesc> arg_types; |
156 | 0 | arg_types.push_back(create_type_desc(slot_data_type->get_primitive_type(), |
157 | 0 | slot_data_type->get_precision(), |
158 | 0 | slot_data_type->get_scale())); |
159 | |
|
160 | 0 | arg_types.push_back(create_type_desc(slot_data_type->get_primitive_type(), |
161 | 0 | slot_data_type->get_precision(), |
162 | 0 | slot_data_type->get_scale())); |
163 | 0 | fn.__set_arg_types(arg_types); |
164 | 0 | fn.__set_ret_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN)); |
165 | 0 | fn.__set_has_var_args(false); |
166 | |
|
167 | 0 | TExprNode texpr_node; |
168 | 0 | texpr_node.__set_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN)); |
169 | 0 | texpr_node.__set_node_type(TExprNodeType::BINARY_PRED); |
170 | 0 | texpr_node.__set_opcode(_predicate->is_asc() ? TExprOpcode::LE : TExprOpcode::GE); |
171 | 0 | texpr_node.__set_fn(fn); |
172 | 0 | texpr_node.__set_num_children(2); |
173 | 0 | texpr_node.__set_is_nullable(is_nullable()); |
174 | 0 | new_root = VectorizedFnCall::create_shared(texpr_node); |
175 | 0 | } |
176 | |
|
177 | 0 | { |
178 | | // add slot |
179 | 0 | new_root->add_child(children().at(0)); |
180 | 0 | } |
181 | | // add Literal |
182 | 0 | { |
183 | 0 | Field field = _predicate->get_value(); |
184 | 0 | TExprNode node = create_texpr_node_from(field, slot_data_type->get_primitive_type(), |
185 | 0 | slot_data_type->get_precision(), |
186 | 0 | slot_data_type->get_scale()); |
187 | 0 | new_root->add_child(VLiteral::create_shared(node)); |
188 | 0 | } |
189 | | |
190 | | // Since the normal greater than or less than relationship does not consider the relationship of null values, the generated `col >=/<= xxx OR col is null.` |
191 | 0 | if (_predicate->nulls_first()) { |
192 | 0 | VExprSPtr col_is_null_node; |
193 | 0 | { |
194 | 0 | TFunction fn; |
195 | 0 | TFunctionName fn_name; |
196 | 0 | fn_name.__set_db_name(""); |
197 | 0 | fn_name.__set_function_name("is_null_pred"); |
198 | 0 | fn.__set_name(fn_name); |
199 | 0 | fn.__set_binary_type(TFunctionBinaryType::BUILTIN); |
200 | 0 | std::vector<TTypeDesc> arg_types; |
201 | 0 | arg_types.push_back(create_type_desc(slot_data_type->get_primitive_type(), |
202 | 0 | slot_data_type->get_precision(), |
203 | 0 | slot_data_type->get_scale())); |
204 | 0 | fn.__set_arg_types(arg_types); |
205 | 0 | fn.__set_ret_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN)); |
206 | 0 | fn.__set_has_var_args(false); |
207 | |
|
208 | 0 | TExprNode texpr_node; |
209 | 0 | texpr_node.__set_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN)); |
210 | 0 | texpr_node.__set_node_type(TExprNodeType::FUNCTION_CALL); |
211 | 0 | texpr_node.__set_fn(fn); |
212 | 0 | texpr_node.__set_num_children(1); |
213 | 0 | col_is_null_node = VectorizedFnCall::create_shared(texpr_node); |
214 | | |
215 | | // add slot. |
216 | 0 | col_is_null_node->add_child(children().at(0)); |
217 | 0 | } |
218 | |
|
219 | 0 | VExprSPtr or_node; |
220 | 0 | { |
221 | 0 | TExprNode texpr_node; |
222 | 0 | texpr_node.__set_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN)); |
223 | 0 | texpr_node.__set_node_type(TExprNodeType::COMPOUND_PRED); |
224 | 0 | texpr_node.__set_opcode(TExprOpcode::COMPOUND_OR); |
225 | 0 | texpr_node.__set_num_children(2); |
226 | 0 | or_node = VectorizedFnCall::create_shared(texpr_node); |
227 | 0 | } |
228 | |
|
229 | 0 | or_node->add_child(col_is_null_node); |
230 | 0 | or_node->add_child(new_root); |
231 | 0 | new_root = or_node; |
232 | 0 | } |
233 | |
|
234 | 0 | return true; |
235 | 0 | } |
236 | | |
237 | | private: |
238 | | int _source_node_id; |
239 | | std::string _expr_name; |
240 | | RuntimePredicate* _predicate = nullptr; |
241 | | FunctionBasePtr _function; |
242 | | VExprContextSPtr _target_ctx; |
243 | | }; |
244 | | |
245 | | #include "common/compile_check_end.h" |
246 | | } // namespace doris |