Coverage Report

Created: 2026-05-08 23:56

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/exprs/vcompound_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
#include <gen_cpp/Opcodes_types.h>
20
21
#include <algorithm>
22
#include <cstdint>
23
24
#include "common/status.h"
25
#include "core/assert_cast.h"
26
#include "core/column/column.h"
27
#include "core/column/column_nullable.h"
28
#include "exprs/vectorized_fn_call.h"
29
#include "exprs/vexpr_context.h"
30
#include "exprs/vexpr_fwd.h"
31
#include "util/simd/bits.h"
32
33
namespace doris {
34
35
17
inline std::string compound_operator_to_string(TExprOpcode::type op) {
36
17
    if (op == TExprOpcode::COMPOUND_AND) {
37
8
        return "and";
38
9
    } else if (op == TExprOpcode::COMPOUND_OR) {
39
9
        return "or";
40
9
    } else {
41
0
        return "not";
42
0
    }
43
17
}
44
45
class VCompoundPred : public VectorizedFnCall {
46
    ENABLE_FACTORY_CREATOR(VCompoundPred);
47
48
public:
49
17
    VCompoundPred(const TExprNode& node) : VectorizedFnCall(node) {
50
17
        _op = node.opcode;
51
17
        _fn.name.function_name = compound_operator_to_string(_op);
52
17
        _expr_name = fmt::format("VCompoundPredicate[{}](arguments={},return={})",
53
17
                                 _fn.name.function_name, get_child_names(), _data_type->get_name());
54
17
    }
55
56
#ifdef BE_TEST
57
2
    VCompoundPred() = default;
58
#endif
59
60
18
    const std::string& expr_name() const override { return _expr_name; }
61
62
0
    Status evaluate_inverted_index(VExprContext* context, uint32_t segment_num_rows) override {
63
0
        segment_v2::InvertedIndexResultBitmap res;
64
0
        bool all_pass = true;
65
66
0
        switch (_op) {
67
0
        case TExprOpcode::COMPOUND_OR: {
68
0
            for (const auto& child : _children) {
69
0
                if (Status st = child->evaluate_inverted_index(context, segment_num_rows);
70
0
                    !st.ok()) {
71
0
                    LOG(ERROR) << "expr:" << child->expr_name()
72
0
                               << " evaluate_inverted_index error:" << st.to_string();
73
0
                    all_pass = false;
74
0
                    continue;
75
0
                }
76
0
                auto inverted_index_context = context->get_index_context();
77
0
                if (inverted_index_context->has_index_result_for_expr(child.get())) {
78
0
                    const auto* index_result =
79
0
                            inverted_index_context->get_index_result_for_expr(child.get());
80
0
                    if (res.is_empty()) {
81
0
                        res = *index_result;
82
0
                    } else {
83
0
                        res |= *index_result;
84
0
                    }
85
0
                    if (inverted_index_context->get_score_runtime() == nullptr) {
86
0
                        if (res.get_data_bitmap()->cardinality() == segment_num_rows) {
87
0
                            break; // Early exit if result is full
88
0
                        }
89
0
                    }
90
0
                } else {
91
0
                    all_pass = false;
92
0
                }
93
0
            }
94
0
            break;
95
0
        }
96
0
        case TExprOpcode::COMPOUND_AND: {
97
0
            for (const auto& child : _children) {
98
0
                if (Status st = child->evaluate_inverted_index(context, segment_num_rows);
99
0
                    !st.ok()) {
100
0
                    LOG(ERROR) << "expr:" << child->expr_name()
101
0
                               << " evaluate_inverted_index error:" << st.to_string();
102
0
                    all_pass = false;
103
0
                    continue;
104
0
                }
105
0
                if (context->get_index_context()->has_index_result_for_expr(child.get())) {
106
0
                    const auto* index_result =
107
0
                            context->get_index_context()->get_index_result_for_expr(child.get());
108
0
                    if (res.is_empty()) {
109
0
                        res = *index_result;
110
0
                    } else {
111
0
                        res &= *index_result;
112
0
                    }
113
114
0
                    if (res.get_data_bitmap()->isEmpty()) {
115
0
                        break; // Early exit if result is empty
116
0
                    }
117
0
                } else {
118
0
                    all_pass = false;
119
0
                }
120
0
            }
121
0
            break;
122
0
        }
123
0
        case TExprOpcode::COMPOUND_NOT: {
124
0
            const auto& child = _children[0];
125
0
            Status st = child->evaluate_inverted_index(context, segment_num_rows);
126
0
            if (!st.ok()) {
127
0
                LOG(ERROR) << "expr:" << child->expr_name()
128
0
                           << " evaluate_inverted_index error:" << st.to_string();
129
0
                return st;
130
0
            }
131
132
0
            if (context->get_index_context()->has_index_result_for_expr(child.get())) {
133
0
                const auto* index_result =
134
0
                        context->get_index_context()->get_index_result_for_expr(child.get());
135
0
                roaring::Roaring full_result;
136
0
                full_result.addRange(0, segment_num_rows);
137
0
                res = index_result->op_not(&full_result);
138
0
            } else {
139
0
                all_pass = false;
140
0
            }
141
0
            break;
142
0
        }
143
0
        default:
144
0
            return Status::NotSupported(
145
0
                    "Compound operator must be AND, OR, or NOT to execute with inverted index.");
146
0
        }
147
148
0
        if (all_pass && !res.is_empty()) {
149
0
            context->get_index_context()->set_index_result_for_expr(this, res);
150
0
        }
151
0
        return Status::OK();
152
0
    }
153
154
    Status execute_column_impl(VExprContext* context, const Block* block, const Selector* selector,
155
0
                               size_t count, ColumnPtr& result_column) const override {
156
0
        if (fast_execute(context, selector, count, result_column)) {
157
0
            return Status::OK();
158
0
        }
159
0
        if (get_num_children() == 1 || _has_const_child()) {
160
0
            return VectorizedFnCall::execute_column_impl(context, block, selector, count,
161
0
                                                         result_column);
162
0
        }
163
164
0
        ColumnPtr lhs_column;
165
0
        RETURN_IF_ERROR(_children[0]->execute_column(context, block, selector, count, lhs_column));
166
0
        lhs_column = lhs_column->convert_to_full_column_if_const();
167
0
        size_t size = lhs_column->size();
168
169
0
        bool lhs_is_nullable = lhs_column->is_nullable();
170
0
        auto [lhs_data_column, lhs_null_map] =
171
0
                _get_raw_data_and_null_map(lhs_column, lhs_is_nullable);
172
0
        size_t filted = simd::count_zero_num((int8_t*)lhs_data_column, size);
173
0
        bool lhs_all_true = (filted == 0);
174
0
        bool lhs_all_false = (filted == size);
175
176
0
        bool lhs_all_is_not_null = false;
177
0
        if (lhs_is_nullable) {
178
0
            filted = simd::count_zero_num((int8_t*)lhs_null_map, size);
179
0
            lhs_all_is_not_null = (filted == size);
180
0
        }
181
182
0
        ColumnPtr rhs_column = nullptr;
183
0
        uint8_t* __restrict rhs_data_column = nullptr;
184
0
        uint8_t* __restrict rhs_null_map = nullptr;
185
0
        bool rhs_is_nullable = false;
186
0
        bool rhs_all_true = false;
187
0
        bool rhs_all_false = false;
188
0
        bool rhs_all_is_not_null = false;
189
0
        bool result_is_nullable = _data_type->is_nullable();
190
191
0
        auto get_rhs_colum = [&]() {
192
0
            if (!rhs_column) {
193
0
                RETURN_IF_ERROR(
194
0
                        _children[1]->execute_column(context, block, selector, count, rhs_column));
195
0
                rhs_column = rhs_column->convert_to_full_column_if_const();
196
0
                rhs_is_nullable = rhs_column->is_nullable();
197
0
                auto rhs_nullable_column = _get_raw_data_and_null_map(rhs_column, rhs_is_nullable);
198
0
                rhs_data_column = rhs_nullable_column.first;
199
0
                rhs_null_map = rhs_nullable_column.second;
200
0
                size_t filted = simd::count_zero_num((int8_t*)rhs_data_column, size);
201
0
                rhs_all_true = (filted == 0);
202
0
                rhs_all_false = (filted == size);
203
0
                if (rhs_is_nullable) {
204
0
                    filted = simd::count_zero_num((int8_t*)rhs_null_map, size);
205
0
                    rhs_all_is_not_null = (filted == size);
206
0
                }
207
0
            }
208
0
            return Status::OK();
209
0
        };
210
211
0
        auto return_result_column_id = [&](ColumnPtr& arg_column) {
212
0
            result_column = std::move(*arg_column).mutate();
213
0
            if (result_is_nullable && !result_column->is_nullable()) {
214
0
                result_column = make_nullable(result_column);
215
0
            }
216
0
        };
217
218
0
        auto create_null_map_column = [&](ColumnPtr& null_map_column,
219
0
                                          uint8_t* __restrict null_map_data) {
220
0
            if (null_map_data == nullptr) {
221
0
                null_map_column = ColumnUInt8::create(size, 0);
222
0
                null_map_data = assert_cast<ColumnUInt8*>(null_map_column->assume_mutable().get())
223
0
                                        ->get_data()
224
0
                                        .data();
225
0
            }
226
0
            return null_map_data;
227
0
        };
228
229
0
        auto vector_vector = [&]<bool is_and_op>() {
230
0
            if (lhs_column->use_count() == 1) {
231
0
                result_column = lhs_column;
232
0
            } else if (rhs_column->use_count() == 1) {
233
0
                result_column = rhs_column;
234
0
                auto tmp_column = rhs_data_column;
235
0
                rhs_data_column = lhs_data_column;
236
0
                lhs_data_column = tmp_column;
237
0
            } else {
238
0
                auto col_res = lhs_column->clone_resized(size);
239
0
                lhs_data_column = assert_cast<ColumnUInt8*>(col_res.get())->get_data().data();
240
0
                result_column = std::move(col_res);
241
0
            }
242
243
0
            do_not_null_pred<is_and_op>(lhs_data_column, rhs_data_column, size);
244
0
        };
Unexecuted instantiation: _ZZNK5doris13VCompoundPred19execute_column_implEPNS_12VExprContextEPKNS_5BlockEPKNS_8PODArrayIjLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEEmRNS_3COWINS_7IColumnEE13immutable_ptrISE_EEENKUlTnbvE_clILb1EEEDav
Unexecuted instantiation: _ZZNK5doris13VCompoundPred19execute_column_implEPNS_12VExprContextEPKNS_5BlockEPKNS_8PODArrayIjLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEEmRNS_3COWINS_7IColumnEE13immutable_ptrISE_EEENKUlTnbvE_clILb0EEEDav
245
0
        auto vector_vector_null = [&]<bool is_and_op>() {
246
0
            auto col_res = ColumnUInt8::create(size);
247
0
            auto col_nulls = ColumnUInt8::create(size);
248
249
0
            auto* __restrict res_datas = col_res->get_data().data();
250
0
            auto* __restrict res_nulls = col_nulls->get_data().data();
251
0
            ColumnPtr temp_null_map = nullptr;
252
            // maybe both children are nullable / or one of children is nullable
253
0
            auto* __restrict lhs_null_map_tmp = create_null_map_column(temp_null_map, lhs_null_map);
254
0
            auto* __restrict rhs_null_map_tmp = create_null_map_column(temp_null_map, rhs_null_map);
255
0
            auto* __restrict lhs_data_column_tmp = lhs_data_column;
256
0
            auto* __restrict rhs_data_column_tmp = rhs_data_column;
257
258
0
            do_null_pred<is_and_op>(lhs_data_column_tmp, lhs_null_map_tmp, rhs_data_column_tmp,
259
0
                                    rhs_null_map_tmp, res_datas, res_nulls, size);
260
261
0
            result_column = ColumnNullable::create(std::move(col_res), std::move(col_nulls));
262
0
        };
Unexecuted instantiation: _ZZNK5doris13VCompoundPred19execute_column_implEPNS_12VExprContextEPKNS_5BlockEPKNS_8PODArrayIjLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEEmRNS_3COWINS_7IColumnEE13immutable_ptrISE_EEENKUlTnbvE0_clILb1EEEDav
Unexecuted instantiation: _ZZNK5doris13VCompoundPred19execute_column_implEPNS_12VExprContextEPKNS_5BlockEPKNS_8PODArrayIjLm4096ENS_9AllocatorILb0ELb0ELb0ENS_22DefaultMemoryAllocatorELb1EEELm16ELm15EEEmRNS_3COWINS_7IColumnEE13immutable_ptrISE_EEENKUlTnbvE0_clILb0EEEDav
263
264
        // false and NULL ----> 0
265
        // true  and NULL ----> NULL
266
0
        if (_op == TExprOpcode::COMPOUND_AND) {
267
            //1. not null column: all data is false
268
            //2. nullable column: null map all is not null
269
0
            if ((lhs_all_false && !lhs_is_nullable) || (lhs_all_false && lhs_all_is_not_null)) {
270
                // false and any = false, return lhs
271
0
                return_result_column_id(lhs_column);
272
0
            } else {
273
0
                RETURN_IF_ERROR(get_rhs_colum());
274
275
0
                if ((lhs_all_true && !lhs_is_nullable) ||    //not null column
276
0
                    (lhs_all_true && lhs_all_is_not_null)) { //nullable column
277
                                                             // true and any = any, return rhs
278
279
0
                    return_result_column_id(rhs_column);
280
0
                } else if ((rhs_all_false && !rhs_is_nullable) ||
281
0
                           (rhs_all_false && rhs_all_is_not_null)) {
282
                    // any and false = false, return rhs
283
0
                    return_result_column_id(rhs_column);
284
0
                } else if ((rhs_all_true && !rhs_is_nullable) ||
285
0
                           (rhs_all_true && rhs_all_is_not_null)) {
286
                    // any and true = any, return lhs
287
0
                    return_result_column_id(lhs_column);
288
0
                } else {
289
0
                    if (!result_is_nullable) {
290
0
                        vector_vector.template operator()<true>();
291
0
                    } else {
292
0
                        vector_vector_null.template operator()<true>();
293
0
                    }
294
0
                }
295
0
            }
296
0
        } else if (_op == TExprOpcode::COMPOUND_OR) {
297
            // true  or NULL ----> 1
298
            // false or NULL ----> NULL
299
0
            if ((lhs_all_true && !lhs_is_nullable) || (lhs_all_true && lhs_all_is_not_null)) {
300
                // true or any = true, return lhs
301
0
                return_result_column_id(lhs_column);
302
0
            } else {
303
0
                RETURN_IF_ERROR(get_rhs_colum());
304
0
                if ((lhs_all_false && !lhs_is_nullable) || (lhs_all_false && lhs_all_is_not_null)) {
305
                    // false or any = any, return rhs
306
0
                    return_result_column_id(rhs_column);
307
0
                } else if ((rhs_all_true && !rhs_is_nullable) ||
308
0
                           (rhs_all_true && rhs_all_is_not_null)) {
309
                    // any or true = true, return rhs
310
0
                    return_result_column_id(rhs_column);
311
0
                } else if ((rhs_all_false && !rhs_is_nullable) ||
312
0
                           (rhs_all_false && rhs_all_is_not_null)) {
313
                    // any or false = any, return lhs
314
0
                    return_result_column_id(lhs_column);
315
0
                } else {
316
0
                    if (!result_is_nullable) {
317
0
                        vector_vector.template operator()<false>();
318
0
                    } else {
319
0
                        vector_vector_null.template operator()<false>();
320
0
                    }
321
0
                }
322
0
            }
323
0
        } else {
324
0
            return Status::InternalError("Compound operator must be AND or OR.");
325
0
        }
326
327
0
        DCHECK_EQ(result_column->size(), count);
328
0
        return Status::OK();
329
0
    }
330
331
0
    double execute_cost() const override {
332
0
        double cost = 0.3;
333
0
        for (const auto& child : _children) {
334
0
            cost += child->execute_cost();
335
0
        }
336
0
        return cost;
337
0
    }
338
339
private:
340
0
    static inline constexpr uint8_t apply_and_null(UInt8 a, UInt8 l_null, UInt8 b, UInt8 r_null) {
341
        // (<> && false) is false, (true && NULL) is NULL
342
0
        return (l_null & r_null) | (r_null & (l_null ^ a)) | (l_null & (r_null ^ b));
343
0
    }
344
0
    static inline constexpr uint8_t apply_or_null(UInt8 a, UInt8 l_null, UInt8 b, UInt8 r_null) {
345
        // (<> || true) is true, (false || NULL) is NULL
346
0
        return (l_null & r_null) | (r_null & (r_null ^ a)) | (l_null & (l_null ^ b));
347
0
    }
348
349
    template <bool is_and>
350
0
    void static do_not_null_pred(uint8_t* __restrict lhs, uint8_t* __restrict rhs, size_t size) {
351
#ifdef NDEBUG
352
#if defined(__clang__)
353
#pragma clang loop vectorize(enable)
354
#elif defined(__GNUC__) && (__GNUC__ >= 5)
355
#pragma GCC ivdep
356
#endif
357
#endif
358
0
        for (size_t i = 0; i < size; ++i) {
359
0
            if constexpr (is_and) {
360
0
                lhs[i] &= rhs[i];
361
0
            } else {
362
0
                lhs[i] |= rhs[i];
363
0
            }
364
0
        }
365
0
    }
Unexecuted instantiation: _ZN5doris13VCompoundPred16do_not_null_predILb1EEEvPhS2_m
Unexecuted instantiation: _ZN5doris13VCompoundPred16do_not_null_predILb0EEEvPhS2_m
366
367
    template <bool is_and>
368
    void static do_null_pred(uint8_t* __restrict lhs_data, uint8_t* __restrict lhs_null,
369
                             uint8_t* __restrict rhs_data, uint8_t* __restrict rhs_null,
370
                             uint8_t* __restrict res_data, uint8_t* __restrict res_null,
371
0
                             size_t size) {
372
#ifdef NDEBUG
373
#if defined(__clang__)
374
#pragma clang loop vectorize(enable)
375
#elif defined(__GNUC__) && (__GNUC__ >= 5)
376
#pragma GCC ivdep
377
#endif
378
#endif
379
0
        for (size_t i = 0; i < size; ++i) {
380
0
            if constexpr (is_and) {
381
0
                res_null[i] = apply_and_null(lhs_data[i], lhs_null[i], rhs_data[i], rhs_null[i]);
382
0
                res_data[i] = lhs_data[i] & rhs_data[i];
383
0
            } else {
384
0
                res_null[i] = apply_or_null(lhs_data[i], lhs_null[i], rhs_data[i], rhs_null[i]);
385
0
                res_data[i] = lhs_data[i] | rhs_data[i];
386
0
            }
387
0
        }
388
0
    }
Unexecuted instantiation: _ZN5doris13VCompoundPred12do_null_predILb1EEEvPhS2_S2_S2_S2_S2_m
Unexecuted instantiation: _ZN5doris13VCompoundPred12do_null_predILb0EEEvPhS2_S2_S2_S2_S2_m
389
390
0
    bool _has_const_child() const {
391
0
        return std::ranges::any_of(_children,
392
0
                                   [](const VExprSPtr& arg) -> bool { return arg->is_constant(); });
393
0
    }
394
395
    std::pair<uint8_t*, uint8_t*> _get_raw_data_and_null_map(ColumnPtr column,
396
0
                                                             bool has_nullable_column) const {
397
0
        if (has_nullable_column) {
398
0
            auto* nullable_column = assert_cast<ColumnNullable*>(column->assume_mutable().get());
399
0
            auto* data_column =
400
0
                    assert_cast<ColumnUInt8*>(nullable_column->get_nested_column_ptr().get())
401
0
                            ->get_data()
402
0
                            .data();
403
0
            auto* null_map =
404
0
                    assert_cast<ColumnUInt8*>(nullable_column->get_null_map_column_ptr().get())
405
0
                            ->get_data()
406
0
                            .data();
407
0
            return std::make_pair(data_column, null_map);
408
0
        } else {
409
0
            auto* data_column =
410
0
                    assert_cast<ColumnUInt8*>(column->assume_mutable().get())->get_data().data();
411
0
            return std::make_pair(data_column, nullptr);
412
0
        }
413
0
    }
414
415
    TExprOpcode::type _op;
416
};
417
418
} // namespace doris