Coverage Report

Created: 2025-07-27 15:37

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/root/doris/be/src/olap/schema.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 "olap/schema.h"
19
20
#include <glog/logging.h>
21
22
#include <boost/iterator/iterator_facade.hpp>
23
#include <ostream>
24
#include <unordered_set>
25
#include <utility>
26
27
#include "common/config.h"
28
#include "olap/olap_common.h"
29
#include "runtime/define_primitive_type.h"
30
#include "util/trace.h"
31
#include "vec/columns/column_array.h"
32
#include "vec/columns/column_dictionary.h"
33
#include "vec/columns/column_map.h"
34
#include "vec/columns/column_nullable.h"
35
#include "vec/columns/column_struct.h"
36
#include "vec/columns/predicate_column.h"
37
#include "vec/core/types.h"
38
#include "vec/data_types/data_type.h"
39
#include "vec/data_types/data_type_factory.hpp"
40
41
namespace doris {
42
43
11.2M
Schema::Schema(const Schema& other) {
44
11.2M
    _copy_from(other);
45
11.2M
}
46
47
0
Schema& Schema::operator=(const Schema& other) {
48
0
    if (this != &other) {
49
0
        _copy_from(other);
50
0
    }
51
0
    return *this;
52
0
}
53
54
11.2M
void Schema::_copy_from(const Schema& other) {
55
11.2M
    _col_ids = other._col_ids;
56
11.2M
    _col_offsets = other._col_offsets;
57
58
11.2M
    _num_key_columns = other._num_key_columns;
59
11.2M
    _schema_size = other._schema_size;
60
61
    // Deep copy _cols
62
    // TODO(lingbin): really need clone?
63
11.2M
    _cols.resize(other._cols.size(), nullptr);
64
17.5M
    for (auto cid : _col_ids) {
65
17.5M
        _cols[cid] = other._cols[cid]->clone();
66
17.5M
    }
67
11.2M
}
68
69
void Schema::_init(const std::vector<TabletColumnPtr>& cols, const std::vector<ColumnId>& col_ids,
70
16.1M
                   size_t num_key_columns) {
71
16.1M
    _col_ids = col_ids;
72
16.1M
    _num_key_columns = num_key_columns;
73
74
16.1M
    _cols.resize(cols.size(), nullptr);
75
16.1M
    _col_offsets.resize(_cols.size(), -1);
76
77
16.1M
    size_t offset = 0;
78
16.1M
    std::unordered_set<uint32_t> col_id_set(col_ids.begin(), col_ids.end());
79
298M
    for (int cid = 0; cid < cols.size(); ++cid) {
80
282M
        if (col_id_set.find(cid) == col_id_set.end()) {
81
197M
            continue;
82
197M
        }
83
84.7M
        _cols[cid] = FieldFactory::create(*cols[cid]);
84
85
84.7M
        _col_offsets[cid] = offset;
86
        // Plus 1 byte for null byte
87
84.7M
        offset += _cols[cid]->size() + 1;
88
84.7M
    }
89
90
16.1M
    _schema_size = offset;
91
16.1M
}
92
93
void Schema::_init(const std::vector<const Field*>& cols, const std::vector<ColumnId>& col_ids,
94
2.80M
                   size_t num_key_columns) {
95
2.80M
    _col_ids = col_ids;
96
2.80M
    _num_key_columns = num_key_columns;
97
98
2.80M
    _cols.resize(cols.size(), nullptr);
99
2.80M
    _col_offsets.resize(_cols.size(), -1);
100
101
2.80M
    size_t offset = 0;
102
2.80M
    std::unordered_set<uint32_t> col_id_set(col_ids.begin(), col_ids.end());
103
12.7M
    for (int cid = 0; cid < cols.size(); ++cid) {
104
9.94M
        if (col_id_set.find(cid) == col_id_set.end()) {
105
0
            continue;
106
0
        }
107
        // TODO(lingbin): is it necessary to clone Field? each SegmentIterator will
108
        // use this func, can we avoid clone?
109
9.94M
        _cols[cid] = cols[cid]->clone();
110
111
9.94M
        _col_offsets[cid] = offset;
112
        // Plus 1 byte for null byte
113
9.94M
        offset += _cols[cid]->size() + 1;
114
9.94M
    }
115
116
2.80M
    _schema_size = offset;
117
2.80M
}
118
119
30.1M
Schema::~Schema() {
120
421M
    for (auto col : _cols) {
121
421M
        delete col;
122
421M
    }
123
30.1M
}
124
125
140M
vectorized::DataTypePtr Schema::get_data_type_ptr(const Field& field) {
126
140M
    return vectorized::DataTypeFactory::instance().create_data_type(field);
127
140M
}
128
129
10.2M
vectorized::IColumn::MutablePtr Schema::get_column_by_field(const Field& field) {
130
10.2M
    return get_data_type_ptr(field)->create_column();
131
10.2M
}
132
133
vectorized::IColumn::MutablePtr Schema::get_predicate_column_ptr(const FieldType& type,
134
                                                                 bool is_nullable,
135
2.58M
                                                                 const ReaderType reader_type) {
136
2.58M
    vectorized::IColumn::MutablePtr ptr = nullptr;
137
2.58M
    switch (type) {
138
364
    case FieldType::OLAP_FIELD_TYPE_BOOL:
139
364
        ptr = doris::vectorized::PredicateColumnType<TYPE_BOOLEAN>::create();
140
364
        break;
141
2.47M
    case FieldType::OLAP_FIELD_TYPE_TINYINT:
142
2.47M
        ptr = doris::vectorized::PredicateColumnType<TYPE_TINYINT>::create();
143
2.47M
        break;
144
1.40k
    case FieldType::OLAP_FIELD_TYPE_SMALLINT:
145
1.40k
        ptr = doris::vectorized::PredicateColumnType<TYPE_SMALLINT>::create();
146
1.40k
        break;
147
38.9k
    case FieldType::OLAP_FIELD_TYPE_INT:
148
38.9k
        ptr = doris::vectorized::PredicateColumnType<TYPE_INT>::create();
149
38.9k
        break;
150
0
    case FieldType::OLAP_FIELD_TYPE_FLOAT:
151
0
        ptr = doris::vectorized::PredicateColumnType<TYPE_FLOAT>::create();
152
0
        break;
153
0
    case FieldType::OLAP_FIELD_TYPE_DOUBLE:
154
0
        ptr = doris::vectorized::PredicateColumnType<TYPE_DOUBLE>::create();
155
0
        break;
156
12.3k
    case FieldType::OLAP_FIELD_TYPE_BIGINT:
157
12.3k
        ptr = doris::vectorized::PredicateColumnType<TYPE_BIGINT>::create();
158
12.3k
        break;
159
1.29k
    case FieldType::OLAP_FIELD_TYPE_LARGEINT:
160
1.29k
        ptr = doris::vectorized::PredicateColumnType<TYPE_LARGEINT>::create();
161
1.29k
        break;
162
734
    case FieldType::OLAP_FIELD_TYPE_DATE:
163
734
        ptr = doris::vectorized::PredicateColumnType<TYPE_DATE>::create();
164
734
        break;
165
7.43k
    case FieldType::OLAP_FIELD_TYPE_DATEV2:
166
7.43k
        ptr = doris::vectorized::PredicateColumnType<TYPE_DATEV2>::create();
167
7.43k
        break;
168
4.11k
    case FieldType::OLAP_FIELD_TYPE_DATETIMEV2:
169
4.11k
        ptr = doris::vectorized::PredicateColumnType<TYPE_DATETIMEV2>::create();
170
4.11k
        break;
171
376
    case FieldType::OLAP_FIELD_TYPE_DATETIME:
172
376
        ptr = doris::vectorized::PredicateColumnType<TYPE_DATETIME>::create();
173
376
        break;
174
3.34k
    case FieldType::OLAP_FIELD_TYPE_CHAR:
175
3.34k
        if (config::enable_low_cardinality_optimize && reader_type == ReaderType::READER_QUERY) {
176
3.33k
            ptr = doris::vectorized::ColumnDictI32::create(type);
177
3.33k
        } else {
178
10
            ptr = doris::vectorized::PredicateColumnType<TYPE_CHAR>::create();
179
10
        }
180
3.34k
        break;
181
0
    case FieldType::OLAP_FIELD_TYPE_VARCHAR:
182
31.0k
    case FieldType::OLAP_FIELD_TYPE_STRING:
183
31.2k
    case FieldType::OLAP_FIELD_TYPE_JSONB:
184
31.2k
        if (config::enable_low_cardinality_optimize && reader_type == ReaderType::READER_QUERY) {
185
29.2k
            ptr = doris::vectorized::ColumnDictI32::create(type);
186
29.2k
        } else {
187
1.92k
            ptr = doris::vectorized::PredicateColumnType<TYPE_STRING>::create();
188
1.92k
        }
189
31.2k
        break;
190
10
    case FieldType::OLAP_FIELD_TYPE_DECIMAL:
191
10
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMALV2>::create();
192
10
        break;
193
216
    case FieldType::OLAP_FIELD_TYPE_DECIMAL32:
194
216
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMAL32>::create();
195
216
        break;
196
628
    case FieldType::OLAP_FIELD_TYPE_DECIMAL64:
197
628
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMAL64>::create();
198
628
        break;
199
806
    case FieldType::OLAP_FIELD_TYPE_DECIMAL128I:
200
806
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMAL128I>::create();
201
806
        break;
202
388
    case FieldType::OLAP_FIELD_TYPE_DECIMAL256:
203
388
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMAL256>::create();
204
388
        break;
205
270
    case FieldType::OLAP_FIELD_TYPE_IPV4:
206
270
        ptr = doris::vectorized::PredicateColumnType<TYPE_IPV4>::create();
207
270
        break;
208
248
    case FieldType::OLAP_FIELD_TYPE_IPV6:
209
248
        ptr = doris::vectorized::PredicateColumnType<TYPE_IPV6>::create();
210
248
        break;
211
0
    default:
212
0
        throw Exception(
213
0
                ErrorCode::SCHEMA_SCHEMA_FIELD_INVALID,
214
0
                fmt::format("Unexpected type when choosing predicate column, type={}", int(type)));
215
2.58M
    }
216
217
2.58M
    if (is_nullable) {
218
48.2k
        return doris::vectorized::ColumnNullable::create(std::move(ptr),
219
48.2k
                                                         doris::vectorized::ColumnUInt8::create());
220
48.2k
    }
221
2.53M
    return ptr;
222
2.58M
}
223
224
} // namespace doris