Coverage Report

Created: 2025-10-23 11:46

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
5.03M
Schema::Schema(const Schema& other) {
44
5.03M
    _copy_from(other);
45
5.03M
}
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
5.08M
void Schema::_copy_from(const Schema& other) {
55
5.08M
    _col_ids = other._col_ids;
56
5.08M
    _col_offsets = other._col_offsets;
57
58
5.08M
    _num_key_columns = other._num_key_columns;
59
5.08M
    _schema_size = other._schema_size;
60
61
    // Deep copy _cols
62
    // TODO(lingbin): really need clone?
63
5.08M
    _cols.resize(other._cols.size(), nullptr);
64
8.48M
    for (auto cid : _col_ids) {
65
8.48M
        _cols[cid] = other._cols[cid]->clone();
66
8.48M
    }
67
5.08M
}
68
69
void Schema::_init(const std::vector<TabletColumnPtr>& cols, const std::vector<ColumnId>& col_ids,
70
7.77M
                   size_t num_key_columns) {
71
7.77M
    _col_ids = col_ids;
72
7.77M
    _num_key_columns = num_key_columns;
73
74
7.77M
    _cols.resize(cols.size(), nullptr);
75
7.77M
    _col_offsets.resize(_cols.size(), -1);
76
77
7.77M
    size_t offset = 0;
78
7.77M
    std::unordered_set<uint32_t> col_id_set(col_ids.begin(), col_ids.end());
79
116M
    for (int cid = 0; cid < cols.size(); ++cid) {
80
109M
        if (col_id_set.find(cid) == col_id_set.end()) {
81
59.6M
            continue;
82
59.6M
        }
83
49.3M
        _cols[cid] = FieldFactory::create(*cols[cid]);
84
85
49.3M
        _col_offsets[cid] = offset;
86
        // Plus 1 byte for null byte
87
49.3M
        offset += _cols[cid]->size() + 1;
88
49.3M
    }
89
90
7.77M
    _schema_size = offset;
91
7.77M
}
92
93
void Schema::_init(const std::vector<const Field*>& cols, const std::vector<ColumnId>& col_ids,
94
1.77M
                   size_t num_key_columns) {
95
1.77M
    _col_ids = col_ids;
96
1.77M
    _num_key_columns = num_key_columns;
97
98
1.77M
    _cols.resize(cols.size(), nullptr);
99
1.77M
    _col_offsets.resize(_cols.size(), -1);
100
101
1.77M
    size_t offset = 0;
102
1.77M
    std::unordered_set<uint32_t> col_id_set(col_ids.begin(), col_ids.end());
103
8.24M
    for (int cid = 0; cid < cols.size(); ++cid) {
104
6.47M
        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
6.47M
        _cols[cid] = cols[cid]->clone();
110
111
6.47M
        _col_offsets[cid] = offset;
112
        // Plus 1 byte for null byte
113
6.47M
        offset += _cols[cid]->size() + 1;
114
6.47M
    }
115
116
1.77M
    _schema_size = offset;
117
1.77M
}
118
119
14.7M
Schema::~Schema() {
120
174M
    for (auto col : _cols) {
121
174M
        delete col;
122
174M
    }
123
14.7M
}
124
125
66.3M
vectorized::DataTypePtr Schema::get_data_type_ptr(const Field& field) {
126
66.3M
    return vectorized::DataTypeFactory::instance().create_data_type(field);
127
66.3M
}
128
129
6.64M
vectorized::IColumn::MutablePtr Schema::get_column_by_field(const Field& field) {
130
6.64M
    return get_data_type_ptr(field)->create_column();
131
6.64M
}
132
133
vectorized::IColumn::MutablePtr Schema::get_predicate_column_ptr(const FieldType& type,
134
                                                                 bool is_nullable,
135
1.71M
                                                                 const ReaderType reader_type) {
136
1.71M
    vectorized::IColumn::MutablePtr ptr = nullptr;
137
1.71M
    switch (type) {
138
193
    case FieldType::OLAP_FIELD_TYPE_BOOL:
139
193
        ptr = doris::vectorized::PredicateColumnType<TYPE_BOOLEAN>::create();
140
193
        break;
141
1.67M
    case FieldType::OLAP_FIELD_TYPE_TINYINT:
142
1.67M
        ptr = doris::vectorized::PredicateColumnType<TYPE_TINYINT>::create();
143
1.67M
        break;
144
732
    case FieldType::OLAP_FIELD_TYPE_SMALLINT:
145
732
        ptr = doris::vectorized::PredicateColumnType<TYPE_SMALLINT>::create();
146
732
        break;
147
18.0k
    case FieldType::OLAP_FIELD_TYPE_INT:
148
18.0k
        ptr = doris::vectorized::PredicateColumnType<TYPE_INT>::create();
149
18.0k
        break;
150
20
    case FieldType::OLAP_FIELD_TYPE_FLOAT:
151
20
        ptr = doris::vectorized::PredicateColumnType<TYPE_FLOAT>::create();
152
20
        break;
153
118
    case FieldType::OLAP_FIELD_TYPE_DOUBLE:
154
118
        ptr = doris::vectorized::PredicateColumnType<TYPE_DOUBLE>::create();
155
118
        break;
156
4.16k
    case FieldType::OLAP_FIELD_TYPE_BIGINT:
157
4.16k
        ptr = doris::vectorized::PredicateColumnType<TYPE_BIGINT>::create();
158
4.16k
        break;
159
401
    case FieldType::OLAP_FIELD_TYPE_LARGEINT:
160
401
        ptr = doris::vectorized::PredicateColumnType<TYPE_LARGEINT>::create();
161
401
        break;
162
336
    case FieldType::OLAP_FIELD_TYPE_DATE:
163
336
        ptr = doris::vectorized::PredicateColumnType<TYPE_DATE>::create();
164
336
        break;
165
3.29k
    case FieldType::OLAP_FIELD_TYPE_DATEV2:
166
3.29k
        ptr = doris::vectorized::PredicateColumnType<TYPE_DATEV2>::create();
167
3.29k
        break;
168
2.07k
    case FieldType::OLAP_FIELD_TYPE_DATETIMEV2:
169
2.07k
        ptr = doris::vectorized::PredicateColumnType<TYPE_DATETIMEV2>::create();
170
2.07k
        break;
171
397
    case FieldType::OLAP_FIELD_TYPE_DATETIME:
172
397
        ptr = doris::vectorized::PredicateColumnType<TYPE_DATETIME>::create();
173
397
        break;
174
1.76k
    case FieldType::OLAP_FIELD_TYPE_CHAR:
175
1.77k
        if (config::enable_low_cardinality_optimize && reader_type == ReaderType::READER_QUERY) {
176
1.76k
            ptr = doris::vectorized::ColumnDictI32::create(type);
177
1.76k
        } else {
178
7
            ptr = doris::vectorized::PredicateColumnType<TYPE_CHAR>::create();
179
7
        }
180
1.76k
        break;
181
0
    case FieldType::OLAP_FIELD_TYPE_VARCHAR:
182
13.0k
    case FieldType::OLAP_FIELD_TYPE_STRING:
183
13.1k
    case FieldType::OLAP_FIELD_TYPE_JSONB:
184
13.1k
        if (config::enable_low_cardinality_optimize && reader_type == ReaderType::READER_QUERY) {
185
12.1k
            ptr = doris::vectorized::ColumnDictI32::create(type);
186
12.1k
        } else {
187
1.00k
            ptr = doris::vectorized::PredicateColumnType<TYPE_STRING>::create();
188
1.00k
        }
189
13.1k
        break;
190
5
    case FieldType::OLAP_FIELD_TYPE_DECIMAL:
191
5
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMALV2>::create();
192
5
        break;
193
144
    case FieldType::OLAP_FIELD_TYPE_DECIMAL32:
194
144
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMAL32>::create();
195
144
        break;
196
318
    case FieldType::OLAP_FIELD_TYPE_DECIMAL64:
197
318
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMAL64>::create();
198
318
        break;
199
707
    case FieldType::OLAP_FIELD_TYPE_DECIMAL128I:
200
707
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMAL128I>::create();
201
707
        break;
202
193
    case FieldType::OLAP_FIELD_TYPE_DECIMAL256:
203
193
        ptr = doris::vectorized::PredicateColumnType<TYPE_DECIMAL256>::create();
204
193
        break;
205
81
    case FieldType::OLAP_FIELD_TYPE_IPV4:
206
81
        ptr = doris::vectorized::PredicateColumnType<TYPE_IPV4>::create();
207
81
        break;
208
83
    case FieldType::OLAP_FIELD_TYPE_IPV6:
209
83
        ptr = doris::vectorized::PredicateColumnType<TYPE_IPV6>::create();
210
83
        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
1.71M
    }
216
217
1.71M
    if (is_nullable) {
218
23.7k
        return doris::vectorized::ColumnNullable::create(std::move(ptr),
219
23.7k
                                                         doris::vectorized::ColumnUInt8::create());
220
23.7k
    }
221
1.69M
    return ptr;
222
1.71M
}
223
224
} // namespace doris