Coverage Report

Created: 2026-03-11 11:07

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/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
68
0
    Status prepare(RuntimeState* state, const RowDescriptor& desc, VExprContext* context) override {
69
0
        _predicate = &state->get_query_ctx()->get_runtime_predicate(_source_node_id);
70
0
        RETURN_IF_ERROR_OR_PREPARED(VExpr::prepare(state, desc, context));
71
72
0
        ColumnsWithTypeAndName argument_template;
73
0
        argument_template.emplace_back(nullptr, _children[0]->data_type(),
74
0
                                       _children[0]->expr_name());
75
0
        argument_template.emplace_back(nullptr, _children[0]->data_type(), "topn value");
76
77
0
        _function = SimpleFunctionFactory::instance().get_function(
78
0
                _predicate->is_asc() ? "le" : "ge", argument_template, _data_type, {},
79
0
                state->be_exec_version());
80
0
        if (!_function) {
81
0
            return Status::InternalError("get function failed");
82
0
        }
83
0
        return Status::OK();
84
0
    }
85
86
    Status execute_column(VExprContext* context, const Block* block, Selector* selector,
87
0
                          size_t count, ColumnPtr& result_column) const override {
88
0
        if (!_predicate->has_value()) {
89
0
            result_column = create_always_true_column(count, _data_type->is_nullable());
90
0
            return Status::OK();
91
0
        }
92
93
0
        Block temp_block;
94
95
        // slot
96
0
        ColumnPtr slot_column;
97
0
        RETURN_IF_ERROR(_children[0]->execute_column(context, block, selector, count, slot_column));
98
0
        auto slot_type = _children[0]->execute_type(block);
99
0
        temp_block.insert({slot_column, slot_type, _children[0]->expr_name()});
100
0
        int slot_id = 0;
101
102
        // topn value
103
0
        Field field = _predicate->get_value();
104
0
        auto column_ptr = _children[0]->data_type()->create_column_const(1, field);
105
0
        int topn_value_id = VExpr::insert_param(&temp_block,
106
0
                                                {column_ptr, _children[0]->data_type(), _expr_name},
107
0
                                                std::max(count, column_ptr->size()));
108
109
        // if error(slot_id == -1), will return.
110
0
        ColumnNumbers arguments = {static_cast<uint32_t>(slot_id),
111
0
                                   static_cast<uint32_t>(topn_value_id)};
112
113
0
        uint32_t num_columns_without_result = temp_block.columns();
114
        // prepare a column to save result
115
0
        temp_block.insert({nullptr, _data_type, _expr_name});
116
117
0
        RETURN_IF_ERROR(_function->execute(nullptr, temp_block, arguments,
118
0
                                           num_columns_without_result, temp_block.rows()));
119
0
        result_column = std::move(temp_block.get_by_position(num_columns_without_result).column);
120
0
        if (is_nullable() && _predicate->nulls_first()) {
121
            // null values ​​are always not filtered
122
0
            change_null_to_true(result_column->assume_mutable());
123
0
        }
124
0
        DCHECK_EQ(result_column->size(), count);
125
0
        return Status::OK();
126
0
    }
127
128
0
    const std::string& expr_name() const override { return _expr_name; }
129
130
    // only used in external table (for min-max filter). get `slot > xxx`, not `function(slot) > xxx`.
131
0
    bool get_binary_expr(VExprSPtr& new_root) const {
132
0
        if (!get_child(0)->is_slot_ref()) {
133
            // top rf maybe is `xxx order by abs(column) limit xxx`.
134
0
            return false;
135
0
        }
136
137
0
        if (!_predicate->has_value()) {
138
0
            return false;
139
0
        }
140
141
0
        auto* slot_ref = assert_cast<VSlotRef*>(get_child(0).get());
142
0
        auto slot_data_type = remove_nullable(slot_ref->data_type());
143
0
        {
144
0
            TFunction fn;
145
0
            TFunctionName fn_name;
146
0
            fn_name.__set_db_name("");
147
0
            fn_name.__set_function_name(_predicate->is_asc() ? "le" : "ge");
148
0
            fn.__set_name(fn_name);
149
0
            fn.__set_binary_type(TFunctionBinaryType::BUILTIN);
150
0
            std::vector<TTypeDesc> arg_types;
151
0
            arg_types.push_back(create_type_desc(slot_data_type->get_primitive_type(),
152
0
                                                 slot_data_type->get_precision(),
153
0
                                                 slot_data_type->get_scale()));
154
155
0
            arg_types.push_back(create_type_desc(slot_data_type->get_primitive_type(),
156
0
                                                 slot_data_type->get_precision(),
157
0
                                                 slot_data_type->get_scale()));
158
0
            fn.__set_arg_types(arg_types);
159
0
            fn.__set_ret_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN));
160
0
            fn.__set_has_var_args(false);
161
162
0
            TExprNode texpr_node;
163
0
            texpr_node.__set_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN));
164
0
            texpr_node.__set_node_type(TExprNodeType::BINARY_PRED);
165
0
            texpr_node.__set_opcode(_predicate->is_asc() ? TExprOpcode::LE : TExprOpcode::GE);
166
0
            texpr_node.__set_fn(fn);
167
0
            texpr_node.__set_num_children(2);
168
0
            texpr_node.__set_is_nullable(is_nullable());
169
0
            new_root = VectorizedFnCall::create_shared(texpr_node);
170
0
        }
171
172
0
        {
173
            // add slot
174
0
            new_root->add_child(children().at(0));
175
0
        }
176
        // add Literal
177
0
        {
178
0
            Field field = _predicate->get_value();
179
0
            TExprNode node = create_texpr_node_from(field, slot_data_type->get_primitive_type(),
180
0
                                                    slot_data_type->get_precision(),
181
0
                                                    slot_data_type->get_scale());
182
0
            new_root->add_child(VLiteral::create_shared(node));
183
0
        }
184
185
        // 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.`
186
0
        if (_predicate->nulls_first()) {
187
0
            VExprSPtr col_is_null_node;
188
0
            {
189
0
                TFunction fn;
190
0
                TFunctionName fn_name;
191
0
                fn_name.__set_db_name("");
192
0
                fn_name.__set_function_name("is_null_pred");
193
0
                fn.__set_name(fn_name);
194
0
                fn.__set_binary_type(TFunctionBinaryType::BUILTIN);
195
0
                std::vector<TTypeDesc> arg_types;
196
0
                arg_types.push_back(create_type_desc(slot_data_type->get_primitive_type(),
197
0
                                                     slot_data_type->get_precision(),
198
0
                                                     slot_data_type->get_scale()));
199
0
                fn.__set_arg_types(arg_types);
200
0
                fn.__set_ret_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN));
201
0
                fn.__set_has_var_args(false);
202
203
0
                TExprNode texpr_node;
204
0
                texpr_node.__set_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN));
205
0
                texpr_node.__set_node_type(TExprNodeType::FUNCTION_CALL);
206
0
                texpr_node.__set_fn(fn);
207
0
                texpr_node.__set_num_children(1);
208
0
                col_is_null_node = VectorizedFnCall::create_shared(texpr_node);
209
210
                // add slot.
211
0
                col_is_null_node->add_child(children().at(0));
212
0
            }
213
214
0
            VExprSPtr or_node;
215
0
            {
216
0
                TExprNode texpr_node;
217
0
                texpr_node.__set_type(create_type_desc(PrimitiveType::TYPE_BOOLEAN));
218
0
                texpr_node.__set_node_type(TExprNodeType::COMPOUND_PRED);
219
0
                texpr_node.__set_opcode(TExprOpcode::COMPOUND_OR);
220
0
                texpr_node.__set_num_children(2);
221
0
                or_node = VectorizedFnCall::create_shared(texpr_node);
222
0
            }
223
224
0
            or_node->add_child(col_is_null_node);
225
0
            or_node->add_child(new_root);
226
0
            new_root = or_node;
227
0
        }
228
229
0
        return true;
230
0
    }
231
232
private:
233
    int _source_node_id;
234
    std::string _expr_name;
235
    RuntimePredicate* _predicate = nullptr;
236
    FunctionBasePtr _function;
237
    VExprContextSPtr _target_ctx;
238
};
239
240
#include "common/compile_check_end.h"
241
} // namespace doris