Coverage Report

Created: 2026-03-12 16:03

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/exprs/table_function/vexplode_v2.cpp
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
#include "exprs/table_function/vexplode_v2.h"
19
20
#include <glog/logging.h>
21
22
#include <algorithm>
23
#include <cstdint>
24
#include <ostream>
25
26
#include "common/status.h"
27
#include "core/assert_cast.h"
28
#include "core/block/block.h"
29
#include "core/block/column_with_type_and_name.h"
30
#include "core/column/column.h"
31
#include "core/column/column_array.h"
32
#include "core/column/column_nothing.h"
33
#include "core/column/column_variant.h"
34
#include "core/data_type/data_type.h"
35
#include "core/data_type/data_type_array.h"
36
#include "core/data_type/data_type_nothing.h"
37
#include "core/data_type/primitive_type.h"
38
#include "exprs/function/function_helpers.h"
39
#include "exprs/vexpr.h"
40
#include "exprs/vexpr_context.h"
41
42
namespace doris {
43
44
#include "common/compile_check_begin.h"
45
#include "core/column/column_struct.h"
46
47
1.77k
VExplodeV2TableFunction::VExplodeV2TableFunction() {
48
1.77k
    _fn_name = "vexplode";
49
1.77k
}
50
51
Status VExplodeV2TableFunction::_process_init_variant(Block* block, int value_column_idx,
52
23
                                                      int children_column_idx) {
53
    // explode variant array
54
23
    auto column_without_nullable = remove_nullable(block->get_by_position(value_column_idx).column);
55
23
    auto column = column_without_nullable->convert_to_full_column_if_const();
56
23
    auto& variant_column = assert_cast<ColumnVariant&>(*(column->assume_mutable()));
57
23
    variant_column.finalize();
58
23
    _multi_detail[children_column_idx].output_as_variant = true;
59
23
    if (!variant_column.is_null_root()) {
60
22
        _array_columns[children_column_idx] = variant_column.get_root();
61
        // We need to wrap the output nested column within a variant column.
62
        // Otherwise the type is missmatched
63
22
        const auto* array_type = check_and_get_data_type<DataTypeArray>(
64
22
                remove_nullable(variant_column.get_root_type()).get());
65
22
        if (array_type == nullptr) {
66
2
            return Status::NotSupported("explode not support none array type {}",
67
2
                                        variant_column.get_root_type()->get_name());
68
2
        }
69
20
        _multi_detail[children_column_idx].nested_type = array_type->get_nested_type();
70
20
    } else {
71
        // null root, use nothing type
72
1
        _array_columns[children_column_idx] = ColumnNullable::create(
73
1
                ColumnArray::create(ColumnNothing::create(0)), ColumnUInt8::create(0));
74
1
        _array_columns[children_column_idx]->assume_mutable()->insert_many_defaults(
75
1
                variant_column.size());
76
1
        _multi_detail[children_column_idx].nested_type = std::make_shared<DataTypeNothing>();
77
1
    }
78
21
    return Status::OK();
79
23
}
80
81
838
Status VExplodeV2TableFunction::process_init(Block* block, RuntimeState* state) {
82
838
    auto expr_size = _expr_context->root()->children().size();
83
838
    CHECK(expr_size >= 1) << "VExplodeV2TableFunction support one or more child but has "
84
0
                          << expr_size;
85
86
838
    int value_column_idx = -1;
87
838
    _multi_detail.resize(expr_size);
88
838
    _array_offsets.resize(expr_size);
89
838
    _array_columns.resize(expr_size);
90
91
1.79k
    for (int i = 0; i < expr_size; i++) {
92
963
        RETURN_IF_ERROR(_expr_context->root()->children()[i]->execute(_expr_context.get(), block,
93
963
                                                                      &value_column_idx));
94
963
        if (block->get_by_position(value_column_idx).type->get_primitive_type() == TYPE_VARIANT) {
95
23
            RETURN_IF_ERROR(_process_init_variant(block, value_column_idx, i));
96
940
        } else {
97
940
            _array_columns[i] = block->get_by_position(value_column_idx)
98
940
                                        .column->convert_to_full_column_if_const();
99
940
        }
100
961
        if (!extract_column_array_info(*_array_columns[i], _multi_detail[i])) {
101
0
            return Status::NotSupported(
102
0
                    "column type {} not supported now",
103
0
                    block->get_by_position(value_column_idx).column->get_name());
104
0
        }
105
961
    }
106
107
836
    return Status::OK();
108
838
}
109
110
17.2k
void VExplodeV2TableFunction::process_row(size_t row_idx) {
111
17.2k
    TableFunction::process_row(row_idx);
112
113
34.8k
    for (int i = 0; i < _multi_detail.size(); i++) {
114
17.5k
        auto& detail = _multi_detail[i];
115
17.5k
        if (!detail.array_nullmap_data || !detail.array_nullmap_data[row_idx]) {
116
12.3k
            _array_offsets[i] = (*detail.offsets_ptr)[row_idx - 1];
117
            // find max size in array
118
12.3k
            auto cur_size = (*detail.offsets_ptr)[row_idx] - _array_offsets[i];
119
12.3k
            _cur_size = std::max<unsigned long>(_cur_size, cur_size);
120
12.3k
        }
121
17.5k
    }
122
17.2k
    _row_idx = row_idx;
123
17.2k
}
124
125
824
void VExplodeV2TableFunction::process_close() {
126
824
    _multi_detail.clear();
127
824
    _array_offsets.clear();
128
824
    _array_columns.clear();
129
824
    _row_idx = 0;
130
824
}
131
132
31
void VExplodeV2TableFunction::get_same_many_values(MutableColumnPtr& column, int length) {
133
31
    if (current_empty()) {
134
6
        column->insert_many_defaults(length);
135
6
        return;
136
6
    }
137
25
    ColumnStruct* struct_column = nullptr;
138
25
    std::vector<IColumn*> columns;
139
140
25
    const bool multi_sub_columns = _multi_detail.size() > 1 || _generate_row_index;
141
142
25
    if (multi_sub_columns) {
143
9
        if (_is_nullable) {
144
9
            auto* nullable_column = assert_cast<ColumnNullable*>(column.get());
145
9
            struct_column =
146
9
                    assert_cast<ColumnStruct*>(nullable_column->get_nested_column_ptr().get());
147
9
            auto* nullmap_column =
148
9
                    assert_cast<ColumnUInt8*>(nullable_column->get_null_map_column_ptr().get());
149
9
            nullmap_column->insert_many_defaults(length);
150
151
9
        } else {
152
0
            struct_column = assert_cast<ColumnStruct*>(column.get());
153
0
        }
154
155
27
        for (size_t i = 0; i != _multi_detail.size(); ++i) {
156
18
            columns.emplace_back(&struct_column->get_column(i + (_generate_row_index ? 1 : 0)));
157
18
        }
158
16
    } else {
159
16
        columns.push_back(column.get());
160
16
    }
161
162
25
    if (_generate_row_index) {
163
0
        auto& pos_column = assert_cast<ColumnInt32&>(struct_column->get_column(0));
164
0
        pos_column.insert_many_vals(static_cast<int32_t>(_cur_offset), length);
165
0
    }
166
167
59
    for (int i = 0; i < _multi_detail.size(); i++) {
168
34
        auto& detail = _multi_detail[i];
169
34
        size_t pos = _array_offsets[i] + _cur_offset;
170
34
        size_t element_size = _multi_detail[i].array_col->size_at(_row_idx);
171
34
        auto& struct_field = *columns.at(i);
172
34
        if ((detail.array_nullmap_data && detail.array_nullmap_data[_row_idx])) {
173
3
            struct_field.insert_many_defaults(length);
174
31
        } else {
175
31
            auto* nullable_column = assert_cast<ColumnNullable*>(struct_field.get_ptr().get());
176
31
            auto* nullmap_column =
177
31
                    assert_cast<ColumnUInt8*>(nullable_column->get_null_map_column_ptr().get());
178
            // only need to check if the value at position pos is null
179
31
            if (element_size < _cur_offset ||
180
31
                (detail.nested_nullmap_data && detail.nested_nullmap_data[pos])) {
181
3
                nullable_column->insert_many_defaults(length);
182
28
            } else {
183
28
                nullable_column->get_nested_column_ptr()->insert_many_from(*detail.nested_col, pos,
184
28
                                                                           length);
185
28
                nullmap_column->insert_many_defaults(length);
186
28
            }
187
31
        }
188
34
    }
189
25
}
190
191
1.96k
int VExplodeV2TableFunction::get_value(MutableColumnPtr& column, int max_step) {
192
1.96k
    max_step = std::min(max_step, (int)(_cur_size - _cur_offset));
193
1.96k
    const bool multi_sub_columns = _multi_detail.size() > 1 || _generate_row_index;
194
195
1.96k
    ColumnStruct* struct_column = nullptr;
196
1.96k
    std::vector<IColumn*> columns;
197
198
1.96k
    if (current_empty()) {
199
101
        column->insert_default();
200
101
        max_step = 1;
201
1.86k
    } else {
202
1.86k
        if (multi_sub_columns) {
203
478
            if (_is_nullable) {
204
261
                auto* nullable_column = assert_cast<ColumnNullable*>(column.get());
205
261
                struct_column =
206
261
                        assert_cast<ColumnStruct*>(nullable_column->get_nested_column_ptr().get());
207
261
                auto* nullmap_column =
208
261
                        assert_cast<ColumnUInt8*>(nullable_column->get_null_map_column_ptr().get());
209
261
                nullmap_column->insert_many_defaults(max_step);
210
211
261
            } else {
212
217
                struct_column = assert_cast<ColumnStruct*>(column.get());
213
217
            }
214
215
1.29k
            for (size_t i = 0; i != _multi_detail.size(); ++i) {
216
819
                columns.emplace_back(&struct_column->get_column(i + (_generate_row_index ? 1 : 0)));
217
819
            }
218
1.38k
        } else {
219
1.38k
            columns.emplace_back(column.get());
220
1.38k
        }
221
222
1.86k
        if (_generate_row_index) {
223
367
            auto& pos_column = assert_cast<ColumnInt32&>(struct_column->get_column(0));
224
367
            pos_column.insert_range_of_integer(static_cast<int32_t>(_cur_offset),
225
367
                                               static_cast<int32_t>(_cur_offset + max_step));
226
367
        }
227
228
4.07k
        for (int i = 0; i < _multi_detail.size(); i++) {
229
2.20k
            auto& detail = _multi_detail[i];
230
2.20k
            size_t pos = _array_offsets[i] + _cur_offset;
231
2.20k
            size_t element_size = _multi_detail[i].array_col->size_at(_row_idx);
232
2.20k
            auto& struct_field = *columns.at(i);
233
2.20k
            if (detail.array_nullmap_data && detail.array_nullmap_data[_row_idx]) {
234
42
                struct_field.insert_many_defaults(max_step);
235
2.16k
            } else {
236
2.16k
                auto* nullable_column = assert_cast<ColumnNullable*>(struct_field.get_ptr().get());
237
2.16k
                auto* nullmap_column =
238
2.16k
                        assert_cast<ColumnUInt8*>(nullable_column->get_null_map_column_ptr().get());
239
2.16k
                if (element_size >= _cur_offset + max_step) {
240
1.99k
                    nullable_column->get_nested_column_ptr()->insert_range_from(*detail.nested_col,
241
1.99k
                                                                                pos, max_step);
242
1.99k
                    if (detail.nested_nullmap_data) {
243
1.99k
                        size_t old_size = nullmap_column->size();
244
1.99k
                        nullmap_column->resize(old_size + max_step);
245
1.99k
                        memcpy(nullmap_column->get_data().data() + old_size,
246
1.99k
                               detail.nested_nullmap_data + pos, max_step * sizeof(UInt8));
247
1.99k
                    } else {
248
0
                        nullmap_column->insert_many_defaults(max_step);
249
0
                    }
250
1.99k
                } else if (element_size > _cur_offset) {
251
127
                    auto current_insert_num = element_size - _cur_offset;
252
127
                    nullable_column->get_nested_column_ptr()->insert_range_from(
253
127
                            *detail.nested_col, pos, current_insert_num);
254
127
                    if (detail.nested_nullmap_data) {
255
127
                        size_t old_size = nullmap_column->size();
256
127
                        nullmap_column->resize(old_size + current_insert_num);
257
127
                        memcpy(nullmap_column->get_data().data() + old_size,
258
127
                               detail.nested_nullmap_data + pos,
259
127
                               current_insert_num * sizeof(UInt8));
260
127
                    } else {
261
0
                        nullmap_column->insert_many_defaults(current_insert_num);
262
0
                    }
263
127
                    nullable_column->insert_many_defaults(max_step - current_insert_num);
264
127
                } else {
265
48
                    nullable_column->insert_many_defaults(max_step);
266
48
                }
267
2.16k
            }
268
2.20k
        }
269
1.86k
    }
270
271
1.96k
    forward(max_step);
272
1.96k
    return max_step;
273
1.96k
}
274
275
#include "common/compile_check_end.h"
276
277
} // namespace doris