be/src/storage/predicate/shared_predicate.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 <cstdint> |
21 | | #include <memory> |
22 | | |
23 | | #include "common/factory_creator.h" |
24 | | #include "core/column/column_dictionary.h" |
25 | | #include "storage/index/bloom_filter/bloom_filter.h" |
26 | | #include "storage/index/inverted/inverted_index_reader.h" |
27 | | #include "storage/predicate/column_predicate.h" |
28 | | |
29 | | namespace doris { |
30 | | |
31 | | // SharedPredicate only used on topn runtime predicate. |
32 | | // Runtime predicate globally share one predicate, to ensure that updates can be real-time. |
33 | | // At the beginning nested predicate may be nullptr, in which case predicate always returns true. |
34 | | class SharedPredicate final : public ColumnPredicate { |
35 | | ENABLE_FACTORY_CREATOR(SharedPredicate); |
36 | | |
37 | | public: |
38 | | SharedPredicate(uint32_t column_id, std::string col_name) |
39 | 654 | : ColumnPredicate(column_id, col_name, PrimitiveType::INVALID_TYPE), |
40 | 654 | _mtx(std::make_shared<std::shared_mutex>()) {} |
41 | | SharedPredicate(const ColumnPredicate& other) = delete; |
42 | | SharedPredicate(const SharedPredicate& other, uint32_t column_id) |
43 | | : ColumnPredicate(other, column_id), |
44 | | _mtx(std::make_shared<std::shared_mutex>()), |
45 | 0 | _nested(other._nested ? other._nested->clone(column_id) : nullptr) {} |
46 | 654 | ~SharedPredicate() override = default; |
47 | 2.07k | std::string debug_string() const override { |
48 | 2.07k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
49 | 2.07k | fmt::memory_buffer debug_string_buffer; |
50 | 2.07k | fmt::format_to(debug_string_buffer, "SharedPredicate({}, nested={})", |
51 | 2.07k | ColumnPredicate::debug_string(), _nested ? _nested->debug_string() : "null"); |
52 | 2.07k | return fmt::to_string(debug_string_buffer); |
53 | 2.07k | } |
54 | 3.22k | std::shared_ptr<ColumnPredicate> clone(uint32_t column_id) const override { |
55 | | // All scanner thread should share the same SharedPredicate object. |
56 | 3.22k | return std::const_pointer_cast<ColumnPredicate>(shared_from_this()); |
57 | 3.22k | } |
58 | | |
59 | 13.6k | PredicateType type() const override { |
60 | 13.6k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
61 | 13.6k | if (!_nested) { |
62 | | // topn filter is le or ge |
63 | 6.29k | return PredicateType::LE; |
64 | 6.29k | } |
65 | 7.32k | return _nested->type(); |
66 | 13.6k | } |
67 | 2.07k | PrimitiveType primitive_type() const override { |
68 | 2.07k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
69 | 2.07k | if (!_nested) { |
70 | 452 | return PrimitiveType::INVALID_TYPE; |
71 | 452 | } |
72 | 1.61k | return _nested->primitive_type(); |
73 | 2.07k | } |
74 | | |
75 | 2.43k | void set_nested(const std::shared_ptr<ColumnPredicate>& nested) { |
76 | 2.43k | std::unique_lock<std::shared_mutex> lock(*_mtx); |
77 | 2.43k | _nested = nested; |
78 | 2.43k | } |
79 | | |
80 | | // Only reset between executions, after the old scanners have finished. |
81 | 4 | void reset_for_new_execution() { |
82 | 4 | std::unique_lock<std::shared_mutex> lock(*_mtx); |
83 | 4 | _nested.reset(); |
84 | 4 | } |
85 | | |
86 | | Status evaluate(const IndexFieldNameAndTypePair& name_with_type, IndexIterator* iterator, |
87 | 0 | uint32_t num_rows, roaring::Roaring* bitmap) const override { |
88 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
89 | 0 | if (!_nested) { |
90 | 0 | return Status::OK(); |
91 | 0 | } |
92 | 0 | return _nested->evaluate(name_with_type, iterator, num_rows, bitmap); |
93 | 0 | } |
94 | | |
95 | | void evaluate_and(const IColumn& column, const uint16_t* sel, uint16_t size, |
96 | 0 | bool* flags) const override { |
97 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
98 | 0 | if (!_nested) { |
99 | 0 | return; |
100 | 0 | } |
101 | 0 | return _nested->evaluate_and(column, sel, size, flags); |
102 | 0 | } |
103 | | |
104 | | void evaluate_or(const IColumn& column, const uint16_t* sel, uint16_t size, |
105 | 0 | bool* flags) const override { |
106 | 0 | DCHECK(false) << "should not reach here"; |
107 | 0 | } |
108 | | |
109 | 14.8k | bool evaluate_and(const segment_v2::ZoneMap& zone_map) const override { |
110 | 14.8k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
111 | 14.8k | if (!_nested) { |
112 | 5.40k | return ColumnPredicate::evaluate_and(zone_map); |
113 | 5.40k | } |
114 | 9.48k | return _nested->evaluate_and(zone_map); |
115 | 14.8k | } |
116 | | |
117 | 0 | bool evaluate_del(const segment_v2::ZoneMap& zone_map) const override { |
118 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
119 | 0 | if (!_nested) { |
120 | 0 | return ColumnPredicate::evaluate_del(zone_map); |
121 | 0 | } |
122 | 0 | return _nested->evaluate_del(zone_map); |
123 | 0 | } |
124 | | |
125 | 0 | bool evaluate_and(const BloomFilter* bf) const override { |
126 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
127 | 0 | if (!_nested) { |
128 | 0 | return ColumnPredicate::evaluate_and(bf); |
129 | 0 | } |
130 | 0 | return _nested->evaluate_and(bf); |
131 | 0 | } |
132 | | |
133 | 4.96k | bool can_do_bloom_filter(bool ngram) const override { |
134 | 4.96k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
135 | 4.96k | if (!_nested) { |
136 | 2.84k | return ColumnPredicate::can_do_bloom_filter(ngram); |
137 | 2.84k | } |
138 | 2.12k | return _nested->can_do_bloom_filter(ngram); |
139 | 4.96k | } |
140 | | |
141 | 1.97k | void evaluate_vec(const IColumn& column, uint16_t size, bool* flags) const override { |
142 | 1.97k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
143 | 1.97k | if (!_nested) { |
144 | 2.46M | for (uint16_t i = 0; i < size; ++i) { |
145 | 2.46M | flags[i] = true; |
146 | 2.46M | } |
147 | 883 | return; |
148 | 883 | } |
149 | 1.08k | _nested->evaluate_vec(column, size, flags); |
150 | 1.08k | } |
151 | | |
152 | 1.40k | void evaluate_and_vec(const IColumn& column, uint16_t size, bool* flags) const override { |
153 | 1.40k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
154 | 1.40k | if (!_nested) { |
155 | 526 | return; |
156 | 526 | } |
157 | 881 | _nested->evaluate_and_vec(column, size, flags); |
158 | 881 | } |
159 | | |
160 | 0 | std::string get_search_str() const override { |
161 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
162 | 0 | if (!_nested) { |
163 | 0 | DCHECK(false) << "should not reach here"; |
164 | 0 | } |
165 | 0 | return _nested->get_search_str(); |
166 | 0 | } |
167 | | |
168 | 0 | bool evaluate_and(ParquetPredicate::ColumnStat* statistic) const override { |
169 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
170 | 0 | if (!_nested) { |
171 | | // at the begining _nested will be null, so return true. |
172 | 0 | return true; |
173 | 0 | } |
174 | 0 | return _nested->evaluate_and(statistic); |
175 | 0 | } |
176 | | |
177 | | bool evaluate_and(ParquetPredicate::CachedPageIndexStat* statistic, |
178 | 0 | RowRanges* row_ranges) const override { |
179 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
180 | |
|
181 | 0 | if (!_nested) { |
182 | | // at the begining _nested will be null, so return true. |
183 | 0 | row_ranges->add(statistic->row_group_range); |
184 | 0 | return true; |
185 | 0 | } |
186 | 0 | return _nested->evaluate_and(statistic, row_ranges); |
187 | 0 | } |
188 | | |
189 | | private: |
190 | 340 | uint16_t _evaluate_inner(const IColumn& column, uint16_t* sel, uint16_t size) const override { |
191 | 340 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
192 | 340 | if (!_nested) { |
193 | 109 | return size; |
194 | 109 | } |
195 | 231 | return _nested->evaluate(column, sel, size); |
196 | 340 | } |
197 | | |
198 | | mutable std::shared_ptr<std::shared_mutex> _mtx; |
199 | | std::shared_ptr<ColumnPredicate> _nested; |
200 | | }; |
201 | | |
202 | | } //namespace doris |