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 <atomic> |
21 | | #include <cstdint> |
22 | | #include <memory> |
23 | | |
24 | | #include "common/factory_creator.h" |
25 | | #include "core/column/column_dictionary.h" |
26 | | #include "storage/index/bloom_filter/bloom_filter.h" |
27 | | #include "storage/index/inverted/inverted_index_reader.h" |
28 | | #include "storage/predicate/column_predicate.h" |
29 | | |
30 | | namespace doris { |
31 | | |
32 | | // SharedPredicate only used on topn runtime predicate. |
33 | | // Runtime predicate globally share one predicate, to ensure that updates can be real-time. |
34 | | // At the beginning nested predicate may be nullptr, in which case predicate always returns true. |
35 | | class SharedPredicate final : public ColumnPredicate { |
36 | | ENABLE_FACTORY_CREATOR(SharedPredicate); |
37 | | |
38 | | public: |
39 | | SharedPredicate(uint32_t column_id, std::string col_name) |
40 | 789 | : ColumnPredicate(column_id, col_name, PrimitiveType::INVALID_TYPE), |
41 | 789 | _mtx(std::make_shared<std::shared_mutex>()) {} |
42 | | SharedPredicate(const ColumnPredicate& other) = delete; |
43 | | SharedPredicate(const SharedPredicate& other, uint32_t column_id) |
44 | 0 | : ColumnPredicate(other, column_id), _mtx(std::make_shared<std::shared_mutex>()) { |
45 | 0 | const auto state = other.snapshot(); |
46 | 0 | _nested = state.predicate ? state.predicate->clone(column_id) : nullptr; |
47 | 0 | _version.store(state.version, std::memory_order_relaxed); |
48 | 0 | } |
49 | 789 | ~SharedPredicate() override = default; |
50 | 4.62k | std::string debug_string() const override { |
51 | 4.62k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
52 | 4.62k | fmt::memory_buffer debug_string_buffer; |
53 | 4.62k | fmt::format_to(debug_string_buffer, "SharedPredicate({}, nested={})", |
54 | 4.62k | ColumnPredicate::debug_string(), _nested ? _nested->debug_string() : "null"); |
55 | 4.62k | return fmt::to_string(debug_string_buffer); |
56 | 4.62k | } |
57 | 6.78k | std::shared_ptr<ColumnPredicate> clone(uint32_t column_id) const override { |
58 | | // All scanner thread should share the same SharedPredicate object. |
59 | 6.78k | return std::const_pointer_cast<ColumnPredicate>(shared_from_this()); |
60 | 6.78k | } |
61 | | |
62 | 1.44k | bool is_runtime_filter() const override { return true; } |
63 | | |
64 | 31.0k | PredicateType type() const override { |
65 | 31.0k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
66 | 31.0k | if (!_nested) { |
67 | | // topn filter is le or ge |
68 | 15.8k | return PredicateType::LE; |
69 | 15.8k | } |
70 | 15.2k | return _nested->type(); |
71 | 31.0k | } |
72 | 4.63k | PrimitiveType primitive_type() const override { |
73 | 4.63k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
74 | 4.63k | if (!_nested) { |
75 | 1.49k | return PrimitiveType::INVALID_TYPE; |
76 | 1.49k | } |
77 | 3.13k | return _nested->primitive_type(); |
78 | 4.63k | } |
79 | | |
80 | 3.95k | void set_nested(const std::shared_ptr<ColumnPredicate>& nested) { |
81 | 3.95k | std::unique_lock<std::shared_mutex> lock(*_mtx); |
82 | 3.95k | _nested = nested; |
83 | 3.95k | _version.fetch_add(1, std::memory_order_release); |
84 | 3.95k | } |
85 | | |
86 | 28.0k | uint64_t version() const { return _version.load(std::memory_order_acquire); } |
87 | | |
88 | | struct Snapshot { |
89 | | uint64_t version; |
90 | | std::shared_ptr<const ColumnPredicate> predicate; |
91 | | }; |
92 | | |
93 | | // Keep the immutable comparison alive after releasing the lock. A concurrent update must |
94 | | // not pair version 2 with version 1's bound, or hold the producer lock during index I/O. |
95 | 1.91k | Snapshot snapshot() const { |
96 | 1.91k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
97 | 1.91k | return {.version = _version.load(std::memory_order_relaxed), .predicate = _nested}; |
98 | 1.91k | } |
99 | | |
100 | | Status evaluate(const IndexFieldNameAndTypePair& name_with_type, IndexIterator* iterator, |
101 | 12 | uint32_t num_rows, roaring::Roaring* bitmap) const override { |
102 | 12 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
103 | 12 | if (!_nested) { |
104 | 11 | return Status::OK(); |
105 | 11 | } |
106 | 1 | return _nested->evaluate(name_with_type, iterator, num_rows, bitmap); |
107 | 12 | } |
108 | | |
109 | | void evaluate_and(const IColumn& column, const uint16_t* sel, uint16_t size, |
110 | 0 | bool* flags) const override { |
111 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
112 | 0 | if (!_nested) { |
113 | 0 | return; |
114 | 0 | } |
115 | 0 | return _nested->evaluate_and(column, sel, size, flags); |
116 | 0 | } |
117 | | |
118 | | void evaluate_or(const IColumn& column, const uint16_t* sel, uint16_t size, |
119 | 0 | bool* flags) const override { |
120 | 0 | DCHECK(false) << "should not reach here"; |
121 | 0 | } |
122 | | |
123 | 23.9k | bool evaluate_and(const segment_v2::ZoneMap& zone_map) const override { |
124 | 23.9k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
125 | 23.9k | if (!_nested) { |
126 | 11.8k | return ColumnPredicate::evaluate_and(zone_map); |
127 | 11.8k | } |
128 | 12.1k | return _nested->evaluate_and(zone_map); |
129 | 23.9k | } |
130 | | |
131 | 0 | bool evaluate_del(const segment_v2::ZoneMap& zone_map) const override { |
132 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
133 | 0 | if (!_nested) { |
134 | 0 | return ColumnPredicate::evaluate_del(zone_map); |
135 | 0 | } |
136 | 0 | return _nested->evaluate_del(zone_map); |
137 | 0 | } |
138 | | |
139 | 0 | bool evaluate_and(const BloomFilter* bf) const override { |
140 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
141 | 0 | if (!_nested) { |
142 | 0 | return ColumnPredicate::evaluate_and(bf); |
143 | 0 | } |
144 | 0 | return _nested->evaluate_and(bf); |
145 | 0 | } |
146 | | |
147 | 13.4k | bool can_do_bloom_filter(bool ngram) const override { |
148 | 13.4k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
149 | 13.4k | if (!_nested) { |
150 | 8.52k | return ColumnPredicate::can_do_bloom_filter(ngram); |
151 | 8.52k | } |
152 | 4.91k | return _nested->can_do_bloom_filter(ngram); |
153 | 13.4k | } |
154 | | |
155 | 10.0k | void evaluate_vec(const IColumn& column, uint16_t size, bool* flags) const override { |
156 | 10.0k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
157 | 10.0k | if (!_nested) { |
158 | 8.52M | for (uint16_t i = 0; i < size; ++i) { |
159 | 8.52M | flags[i] = true; |
160 | 8.52M | } |
161 | 4.52k | return; |
162 | 4.52k | } |
163 | 5.49k | _nested->evaluate_vec(column, size, flags); |
164 | 5.49k | } |
165 | | |
166 | 1.21k | void evaluate_and_vec(const IColumn& column, uint16_t size, bool* flags) const override { |
167 | 1.21k | std::shared_lock<std::shared_mutex> lock(*_mtx); |
168 | 1.21k | if (!_nested) { |
169 | 580 | return; |
170 | 580 | } |
171 | 630 | _nested->evaluate_and_vec(column, size, flags); |
172 | 630 | } |
173 | | |
174 | 0 | std::string get_search_str() const override { |
175 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
176 | 0 | if (!_nested) { |
177 | 0 | DCHECK(false) << "should not reach here"; |
178 | 0 | } |
179 | 0 | return _nested->get_search_str(); |
180 | 0 | } |
181 | | |
182 | 0 | bool evaluate_and(ParquetPredicate::ColumnStat* statistic) const override { |
183 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
184 | 0 | if (!_nested) { |
185 | | // at the begining _nested will be null, so return true. |
186 | 0 | return true; |
187 | 0 | } |
188 | 0 | return _nested->evaluate_and(statistic); |
189 | 0 | } |
190 | | |
191 | | bool evaluate_and(ParquetPredicate::CachedPageIndexStat* statistic, |
192 | 0 | RowRanges* row_ranges) const override { |
193 | 0 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
194 | |
|
195 | 0 | if (!_nested) { |
196 | | // at the begining _nested will be null, so return true. |
197 | 0 | row_ranges->add(statistic->row_group_range); |
198 | 0 | return true; |
199 | 0 | } |
200 | 0 | return _nested->evaluate_and(statistic, row_ranges); |
201 | 0 | } |
202 | | |
203 | | private: |
204 | 400 | uint16_t _evaluate_inner(const IColumn& column, uint16_t* sel, uint16_t size) const override { |
205 | 400 | std::shared_lock<std::shared_mutex> lock(*_mtx); |
206 | 400 | if (!_nested) { |
207 | 96 | return size; |
208 | 96 | } |
209 | 304 | return _nested->evaluate(column, sel, size); |
210 | 400 | } |
211 | | |
212 | | mutable std::shared_ptr<std::shared_mutex> _mtx; |
213 | | std::shared_ptr<ColumnPredicate> _nested; |
214 | | std::atomic<uint64_t> _version {0}; |
215 | | }; |
216 | | |
217 | | } //namespace doris |