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 <glog/logging.h> |
21 | | |
22 | | #include <algorithm> |
23 | | #include <cstddef> |
24 | | #include <memory> |
25 | | #include <string> |
26 | | #include <unordered_map> |
27 | | #include <utility> |
28 | | #include <vector> |
29 | | |
30 | | #include "common/factory_creator.h" |
31 | | #include "common/status.h" |
32 | | #include "core/block/block.h" |
33 | | #include "core/block/column_with_type_and_name.h" |
34 | | #include "core/column/column.h" |
35 | | #include "exec/runtime_filter/runtime_filter_selectivity.h" |
36 | | #include "exprs/expr_zonemap_filter.h" |
37 | | #include "exprs/function_context.h" |
38 | | #include "exprs/vexpr_fwd.h" |
39 | | #include "runtime/runtime_state.h" |
40 | | #include "storage/index/ann/ann_range_search_runtime.h" |
41 | | #include "storage/index/ann/ann_search_params.h" |
42 | | #include "storage/index/inverted/inverted_index_reader.h" |
43 | | #include "storage/index/zone_map/zonemap_filter_result.h" |
44 | | #include "storage/segment/column_reader.h" |
45 | | |
46 | | namespace doris { |
47 | | class RowDescriptor; |
48 | | class RuntimeState; |
49 | | class ZoneMapEvalContext; |
50 | | } // namespace doris |
51 | | |
52 | | namespace doris::segment_v2 { |
53 | | class Segment; |
54 | | class ColumnIterator; |
55 | | } // namespace doris::segment_v2 |
56 | | |
57 | | namespace doris { |
58 | | |
59 | | class ScoreRuntime; |
60 | | class LambdaExecutionContext; |
61 | | using ScoreRuntimeSPtr = std::shared_ptr<ScoreRuntime>; |
62 | | |
63 | | class IndexExecContext { |
64 | | public: |
65 | | IndexExecContext(const std::vector<std::unique_ptr<segment_v2::IndexIterator>>& index_iterators, |
66 | | const std::vector<IndexFieldNameAndTypePair>& storage_name_and_type_vec, |
67 | | std::unordered_map<ColumnId, std::unordered_map<const VExpr*, bool>>& |
68 | | common_expr_index_status, |
69 | | ScoreRuntimeSPtr score_runtime, segment_v2::Segment* segment, |
70 | | const segment_v2::ColumnIteratorOptions& column_iter_opts) |
71 | 2.11M | : _index_iterators(index_iterators), |
72 | 2.11M | _storage_name_and_type(storage_name_and_type_vec), |
73 | 2.11M | _expr_index_status(common_expr_index_status), |
74 | 2.11M | _score_runtime(std::move(score_runtime)), |
75 | 2.11M | _segment(segment), |
76 | 2.11M | _column_iter_opts(column_iter_opts) {} |
77 | | |
78 | 15.5k | segment_v2::IndexIterator* get_inverted_index_iterator(int32_t read_ordinal) const { |
79 | 15.5k | if (read_ordinal < 0 || static_cast<size_t>(read_ordinal) >= _index_iterators.size()) { |
80 | 0 | return nullptr; |
81 | 0 | } |
82 | 15.5k | const auto& iterator = _index_iterators[read_ordinal]; |
83 | 15.5k | if (!iterator) { |
84 | 3.99k | return nullptr; |
85 | 3.99k | } |
86 | 11.5k | return iterator.get(); |
87 | 15.5k | } |
88 | | |
89 | 16.0k | const IndexFieldNameAndTypePair* get_storage_name_and_type(int32_t read_ordinal) const { |
90 | 16.0k | if (read_ordinal < 0 || |
91 | 16.0k | static_cast<size_t>(read_ordinal) >= _storage_name_and_type.size()) { |
92 | 0 | return nullptr; |
93 | 0 | } |
94 | 16.0k | return &_storage_name_and_type[read_ordinal]; |
95 | 16.0k | } |
96 | | |
97 | 0 | segment_v2::Segment* segment() const { return _segment; } |
98 | | |
99 | 0 | const segment_v2::ColumnIteratorOptions& column_iter_opts() const { return _column_iter_opts; } |
100 | | |
101 | 27.0k | bool has_index_result_for_expr(const VExpr* expr) const { |
102 | 27.0k | return _index_result_bitmap.contains(expr); |
103 | 27.0k | } |
104 | | |
105 | | void set_index_result_for_expr(const VExpr* expr, |
106 | 10.7k | segment_v2::InvertedIndexResultBitmap bitmap) { |
107 | 10.7k | _index_result_bitmap[expr] = std::move(bitmap); |
108 | 10.7k | } |
109 | | |
110 | | std::unordered_map<const VExpr*, segment_v2::InvertedIndexResultBitmap>& |
111 | 17.1k | get_index_result_bitmap() { |
112 | 17.1k | return _index_result_bitmap; |
113 | 17.1k | } |
114 | | |
115 | 28.8k | std::unordered_map<const VExpr*, ColumnPtr>& get_index_result_column() { |
116 | 28.8k | return _index_result_column; |
117 | 28.8k | } |
118 | | |
119 | 10.7k | const segment_v2::InvertedIndexResultBitmap* get_index_result_for_expr(const VExpr* expr) { |
120 | 10.7k | auto iter = _index_result_bitmap.find(expr); |
121 | 10.7k | if (iter == _index_result_bitmap.end()) { |
122 | 0 | return nullptr; |
123 | 0 | } |
124 | 10.7k | return &iter->second; |
125 | 10.7k | } |
126 | | |
127 | 1.17k | void set_index_result_column_for_expr(const VExpr* expr, ColumnPtr column) { |
128 | 1.17k | _index_result_column[expr] = std::move(column); |
129 | 1.17k | } |
130 | | |
131 | 9.39k | void set_true_for_index_status(const VExpr* expr, int32_t read_ordinal) { |
132 | 9.41k | if (_expr_index_status.contains(read_ordinal)) { |
133 | 9.41k | if (_expr_index_status[read_ordinal].contains(expr)) { |
134 | 9.41k | _expr_index_status[read_ordinal][expr] = true; |
135 | 9.41k | } |
136 | 9.41k | } |
137 | 9.39k | } |
138 | | |
139 | 2.97k | ScoreRuntimeSPtr get_score_runtime() const { return _score_runtime; } |
140 | | |
141 | 5.18k | void set_analyzer_ctx_for_expr(const VExpr* expr, InvertedIndexAnalyzerCtxSPtr analyzer_ctx) { |
142 | 5.18k | if (expr == nullptr || analyzer_ctx == nullptr) { |
143 | 0 | return; |
144 | 0 | } |
145 | 5.18k | _expr_analyzer_ctx[expr] = std::move(analyzer_ctx); |
146 | 5.18k | } |
147 | | |
148 | 9.37k | const InvertedIndexAnalyzerCtx* get_analyzer_ctx_for_expr(const VExpr* expr) const { |
149 | 9.37k | auto iter = _expr_analyzer_ctx.find(expr); |
150 | 9.37k | if (iter == _expr_analyzer_ctx.end()) { |
151 | 4.76k | return nullptr; |
152 | 4.76k | } |
153 | 4.60k | return iter->second.get(); |
154 | 9.37k | } |
155 | | |
156 | 2.11M | void set_index_query_context(segment_v2::IndexQueryContextPtr index_query_context) { |
157 | 2.11M | _index_query_context = index_query_context; |
158 | 2.11M | } |
159 | | |
160 | 1.31k | const segment_v2::IndexQueryContextPtr& get_index_query_context() const { |
161 | 1.31k | return _index_query_context; |
162 | 1.31k | } |
163 | | |
164 | | private: |
165 | | // A reference to a vector of unique pointers to index iterators. |
166 | | const std::vector<std::unique_ptr<segment_v2::IndexIterator>>& _index_iterators; |
167 | | |
168 | | // A reference to a vector of storage name and type pairs related to schema. |
169 | | const std::vector<IndexFieldNameAndTypePair>& _storage_name_and_type; |
170 | | |
171 | | // A map of expressions to their corresponding inverted index result bitmaps. |
172 | | std::unordered_map<const VExpr*, segment_v2::InvertedIndexResultBitmap> _index_result_bitmap; |
173 | | |
174 | | // A map of expressions to their corresponding result columns. |
175 | | std::unordered_map<const VExpr*, ColumnPtr> _index_result_column; |
176 | | |
177 | | // Per-expression analyzer context for inverted index evaluation. |
178 | | std::unordered_map<const VExpr*, InvertedIndexAnalyzerCtxSPtr> _expr_analyzer_ctx; |
179 | | |
180 | | // A reference to a map of common expressions to their inverted index evaluation status. |
181 | | std::unordered_map<ColumnId, std::unordered_map<const VExpr*, bool>>& _expr_index_status; |
182 | | |
183 | | ScoreRuntimeSPtr _score_runtime; |
184 | | |
185 | | segment_v2::Segment* _segment = nullptr; // Ref |
186 | | segment_v2::ColumnIteratorOptions _column_iter_opts; |
187 | | segment_v2::IndexQueryContextPtr _index_query_context; |
188 | | }; |
189 | | |
190 | | class VExprContext { |
191 | | ENABLE_FACTORY_CREATOR(VExprContext); |
192 | | |
193 | | public: |
194 | | VExprContext(VExprSPtr expr); |
195 | | ~VExprContext(); |
196 | | [[nodiscard]] Status prepare(RuntimeState* state, const RowDescriptor& row_desc); |
197 | | [[nodiscard]] Status open(RuntimeState* state); |
198 | | [[nodiscard]] Status clone(RuntimeState* state, VExprContextSPtr& new_ctx); |
199 | | [[nodiscard]] Status execute(Block* block, int* result_column_id); |
200 | | [[nodiscard]] Status execute(const Block* block, ColumnPtr& result_column); |
201 | | [[nodiscard]] Status execute(const Block* block, ColumnWithTypeAndName& result_data); |
202 | | [[nodiscard]] DataTypePtr execute_type(const Block* block); |
203 | | [[nodiscard]] const std::string& expr_name() const; |
204 | | [[nodiscard]] bool is_blockable() const; |
205 | | |
206 | | [[nodiscard]] Status execute_const_expr(ColumnWithTypeAndName& result); |
207 | | |
208 | | double execute_cost() const; |
209 | | |
210 | 18.1M | VExprSPtr root() { return _root; } |
211 | 29.8k | void set_root(const VExprSPtr& expr) { _root = expr; } |
212 | 19.9k | void set_index_context(std::shared_ptr<IndexExecContext> index_context) { |
213 | 19.9k | _index_context = std::move(index_context); |
214 | 19.9k | } |
215 | | |
216 | 807k | std::shared_ptr<IndexExecContext> get_index_context() const { return _index_context; } |
217 | | |
218 | | LambdaExecutionContext& lambda_execution_context(); |
219 | | |
220 | | /// Creates a FunctionContext, and returns the index that's passed to fn_context() to |
221 | | /// retrieve the created context. Exprs that need a FunctionContext should call this in |
222 | | /// Prepare() and save the returned index. 'varargs_buffer_size', if specified, is the |
223 | | /// size of the varargs buffer in the created FunctionContext (see udf-internal.h). |
224 | | int register_function_context(RuntimeState* state, const DataTypePtr& return_type, |
225 | | const std::vector<DataTypePtr>& arg_types); |
226 | | |
227 | | /// Retrieves a registered FunctionContext. 'i' is the index returned by the call to |
228 | | /// register_function_context(). This should only be called by VExprs. |
229 | 7.15M | FunctionContext* fn_context(int i) { |
230 | 7.15M | if (i < 0 || i >= _fn_contexts.size()) { |
231 | 0 | throw Exception(ErrorCode::INTERNAL_ERROR, |
232 | 0 | "fn_context index invalid, index={}, _fn_contexts.size()={}", i, |
233 | 0 | _fn_contexts.size()); |
234 | 0 | } |
235 | 7.15M | return _fn_contexts[i].get(); |
236 | 7.15M | } |
237 | | |
238 | | // execute expr with inverted index which column a, b has inverted indexes |
239 | | // but some situation although column b has indexes, but apply index is not useful, we should |
240 | | // skip this expr, just do not apply index anymore. |
241 | | [[nodiscard]] Status evaluate_inverted_index(uint32_t segment_num_rows); |
242 | | |
243 | | [[nodiscard]] static ZoneMapFilterResult evaluate_zonemap_filter( |
244 | | const VExprContextSPtrs& conjuncts, const ZoneMapEvalContext& ctx); |
245 | | [[nodiscard]] static ZoneMapFilterResult evaluate_dictionary_filter( |
246 | | const VExprContextSPtrs& conjuncts, const DictionaryEvalContext& ctx); |
247 | | [[nodiscard]] static ZoneMapFilterResult evaluate_bloom_filter( |
248 | | const VExprContextSPtrs& conjuncts, const BloomFilterEvalContext& ctx); |
249 | | |
250 | | bool all_expr_inverted_index_evaluated(); |
251 | | |
252 | | Status execute_filter(const Block* block, uint8_t* __restrict result_filter_data, size_t rows, |
253 | | bool accept_null, bool* can_filter_all); |
254 | | |
255 | | [[nodiscard]] static Status filter_block(VExprContext* vexpr_ctx, Block* block); |
256 | | |
257 | | [[nodiscard]] static Status filter_block(const VExprContextSPtrs& expr_contexts, Block* block, |
258 | | size_t column_to_keep); |
259 | | |
260 | | [[nodiscard]] static Status execute_conjuncts(const VExprContextSPtrs& ctxs, |
261 | | const std::vector<IColumn::Filter*>* filters, |
262 | | bool accept_null, const Block* block, |
263 | | IColumn::Filter* result_filter, |
264 | | bool* can_filter_all); |
265 | | |
266 | | [[nodiscard]] static Status execute_conjuncts(const VExprContextSPtrs& conjuncts, |
267 | | const Block* block, ColumnUInt8& null_map, |
268 | | IColumn::Filter& result_filter); |
269 | | |
270 | | static Status execute_conjuncts(const VExprContextSPtrs& ctxs, |
271 | | const std::vector<IColumn::Filter*>* filters, Block* block, |
272 | | IColumn::Filter* result_filter, bool* can_filter_all); |
273 | | |
274 | | [[nodiscard]] static Status execute_conjuncts_and_filter_block( |
275 | | const VExprContextSPtrs& ctxs, Block* block, std::vector<uint32_t>& columns_to_filter, |
276 | | int column_to_keep); |
277 | | |
278 | | static Status execute_conjuncts_and_filter_block(const VExprContextSPtrs& ctxs, Block* block, |
279 | | std::vector<uint32_t>& columns_to_filter, |
280 | | int column_to_keep, IColumn::Filter& filter); |
281 | | |
282 | | [[nodiscard]] static Status get_output_block_after_execute_exprs(const VExprContextSPtrs&, |
283 | | const Block&, Block*, |
284 | | bool do_projection = false); |
285 | | |
286 | 17.6k | int get_last_result_column_id() const { |
287 | 17.6k | DCHECK(_last_result_column_id != -1); |
288 | 17.6k | return _last_result_column_id; |
289 | 17.6k | } |
290 | | |
291 | 118k | RuntimeFilterSelectivity& get_runtime_filter_selectivity() { |
292 | 118k | if (!_rf_selectivity) { |
293 | 0 | throw Exception(ErrorCode::INTERNAL_ERROR, "RuntimeFilterSelectivity is null"); |
294 | 0 | } |
295 | 118k | return *_rf_selectivity; |
296 | 118k | } |
297 | | |
298 | 0 | FunctionContext::FunctionStateScope get_function_state_scope() const { |
299 | 0 | return _is_clone ? FunctionContext::THREAD_LOCAL : FunctionContext::FRAGMENT_LOCAL; |
300 | 0 | } |
301 | | |
302 | | void clone_fn_contexts(VExprContext* other); |
303 | | |
304 | | VExprContext& operator=(const VExprContext& other) = delete; |
305 | | |
306 | | VExprContext& operator=(VExprContext&& other) = delete; |
307 | | |
308 | 644k | [[nodiscard]] static size_t get_memory_usage(const VExprContextSPtrs& contexts) { |
309 | 644k | size_t usage = 0; |
310 | 644k | std::for_each(contexts.cbegin(), contexts.cend(), |
311 | 644k | [&usage](auto&& context) { usage += context->_memory_usage; }); |
312 | 644k | return usage; |
313 | 644k | } |
314 | | |
315 | 0 | [[nodiscard]] size_t get_memory_usage() const { return _memory_usage; } |
316 | | |
317 | | void prepare_ann_range_search(const doris::VectorSearchUserParams& params); |
318 | | |
319 | | Status evaluate_ann_range_search( |
320 | | const std::vector<std::unique_ptr<segment_v2::IndexIterator>>& index_iterators, |
321 | | const std::vector<std::unique_ptr<segment_v2::ColumnIterator>>& column_iterators, |
322 | | const std::unordered_map<VExprContext*, std::unordered_map<ColumnId, VExpr*>>& |
323 | | common_expr_to_slotref_map, |
324 | | size_t rows_of_segment, roaring::Roaring& row_bitmap, |
325 | | segment_v2::AnnIndexStats& ann_index_stats, bool enable_result_cache, |
326 | | bool* ann_range_search_executed); |
327 | | |
328 | | uint64_t get_digest(uint64_t seed) const; |
329 | | |
330 | | private: |
331 | | // Close method is called in vexpr context dector, not need call expicility |
332 | | void close(); |
333 | | |
334 | | static void _reset_memory_usage(const VExprContextSPtrs& contexts); |
335 | | |
336 | | friend class VExpr; |
337 | | |
338 | | /// The expr tree this context is for. |
339 | | VExprSPtr _root; |
340 | | |
341 | | /// True if this context came from a Clone() call. Used to manage FunctionStateScope. |
342 | | bool _is_clone = false; |
343 | | |
344 | | /// Variables keeping track of current state. |
345 | | bool _prepared = false; |
346 | | bool _opened = false; |
347 | | |
348 | | /// FunctionContexts for each registered expression. The FunctionContexts are created |
349 | | /// and owned by this VExprContext. |
350 | | std::vector<std::unique_ptr<FunctionContext>> _fn_contexts; |
351 | | |
352 | | int _last_result_column_id = -1; |
353 | | |
354 | | /// The depth of expression-tree. |
355 | | int _depth_num = 0; |
356 | | |
357 | | std::shared_ptr<IndexExecContext> _index_context; |
358 | | size_t _memory_usage = 0; |
359 | | |
360 | | segment_v2::AnnRangeSearchRuntime _ann_range_search_runtime; |
361 | | bool _suitable_for_ann_index = true; |
362 | | |
363 | | std::unique_ptr<LambdaExecutionContext> _lambda_execution_context; |
364 | | |
365 | | std::unique_ptr<RuntimeFilterSelectivity> _rf_selectivity = |
366 | | std::make_unique<RuntimeFilterSelectivity>(); |
367 | | }; |
368 | | } // namespace doris |