Coverage Report

Created: 2026-04-21 19:40

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/runtime/descriptors.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
// This file is copied from
18
// https://github.com/apache/impala/blob/branch-2.9.0/be/src/runtime/descriptors.cc
19
// and modified by Doris
20
21
#include "runtime/descriptors.h"
22
23
#include <fmt/format.h>
24
#include <gen_cpp/Descriptors_types.h>
25
#include <gen_cpp/Types_types.h>
26
#include <gen_cpp/descriptors.pb.h>
27
#include <stddef.h>
28
#include <thrift/protocol/TDebugProtocol.h>
29
30
#include <algorithm>
31
#include <boost/algorithm/string/join.hpp>
32
33
#include "common/exception.h"
34
#include "common/object_pool.h"
35
#include "core/column/column_nothing.h"
36
#include "core/data_type/data_type_array.h"
37
#include "core/data_type/data_type_decimal.h"
38
#include "core/data_type/data_type_factory.hpp"
39
#include "core/data_type/data_type_map.h"
40
#include "core/data_type/data_type_struct.h"
41
#include "core/types.h"
42
#include "exec/common/util.hpp"
43
#include "exprs/aggregate/aggregate_function.h"
44
#include "exprs/function/function_helpers.h"
45
#include "exprs/vexpr.h"
46
#include "util/string_util.h"
47
48
namespace doris {
49
const int RowDescriptor::INVALID_IDX = -1;
50
51
SlotDescriptor::SlotDescriptor(const TSlotDescriptor& tdesc)
52
768k
        : _id(tdesc.id),
53
768k
          _type(DataTypeFactory::instance().create_data_type(tdesc.slotType,
54
768k
                                                             tdesc.nullIndicatorBit != -1)),
55
768k
          _parent(tdesc.parent),
56
768k
          _col_pos(tdesc.columnPos),
57
768k
          _col_name(tdesc.colName),
58
768k
          _col_name_lower_case(to_lower(tdesc.colName)),
59
768k
          _col_unique_id(tdesc.col_unique_id),
60
768k
          _slot_idx(tdesc.slotIdx),
61
768k
          _field_idx(-1),
62
768k
          _is_key(tdesc.is_key),
63
768k
          _column_paths(tdesc.column_paths),
64
768k
          _all_access_paths(tdesc.__isset.all_access_paths ? tdesc.all_access_paths
65
768k
                                                           : TColumnAccessPaths {}),
66
768k
          _predicate_access_paths(tdesc.__isset.predicate_access_paths
67
768k
                                          ? tdesc.predicate_access_paths
68
768k
                                          : TColumnAccessPaths {}),
69
18.4E
          _is_auto_increment(tdesc.__isset.is_auto_increment ? tdesc.is_auto_increment : false),
70
768k
          _col_default_value(tdesc.__isset.col_default_value ? tdesc.col_default_value : "") {
71
768k
    if (tdesc.__isset.virtual_column_expr) {
72
        // Make sure virtual column is valid.
73
100
        if (tdesc.virtual_column_expr.nodes.empty()) {
74
2
            throw doris::Exception(doris::ErrorCode::FATAL_ERROR,
75
2
                                   "Virtual column expr node is empty, col_name: {}, "
76
2
                                   "col_unique_id: {}",
77
2
                                   tdesc.colName, tdesc.col_unique_id);
78
2
        }
79
98
        const auto& node = tdesc.virtual_column_expr.nodes[0];
80
98
        if (node.node_type == TExprNodeType::SLOT_REF) {
81
2
            throw doris::Exception(doris::ErrorCode::FATAL_ERROR,
82
2
                                   "Virtual column expr node is slot ref, col_name: {}, "
83
2
                                   "col_unique_id: {}",
84
2
                                   tdesc.colName, tdesc.col_unique_id);
85
2
        }
86
96
        this->virtual_column_expr = std::make_shared<doris::TExpr>(tdesc.virtual_column_expr);
87
96
    }
88
768k
}
89
90
SlotDescriptor::SlotDescriptor(const PSlotDescriptor& pdesc)
91
84
        : _id(pdesc.id()),
92
84
          _type(DataTypeFactory::instance().create_data_type(pdesc.slot_type(),
93
84
                                                             pdesc.null_indicator_bit() != -1)),
94
84
          _parent(pdesc.parent()),
95
84
          _col_pos(pdesc.column_pos()),
96
84
          _col_name(pdesc.col_name()),
97
84
          _col_name_lower_case(to_lower(pdesc.col_name())),
98
84
          _col_unique_id(pdesc.col_unique_id()),
99
84
          _slot_idx(pdesc.slot_idx()),
100
84
          _field_idx(-1),
101
84
          _is_key(pdesc.is_key()),
102
84
          _column_paths(pdesc.column_paths().begin(), pdesc.column_paths().end()),
103
84
          _is_auto_increment(pdesc.is_auto_increment()) {
104
84
    auto convert_to_thrift_column_access_path = [](const PColumnAccessPath& pb_path) {
105
0
        TColumnAccessPath thrift_path;
106
0
        thrift_path.type = (TAccessPathType::type)pb_path.type();
107
0
        if (pb_path.has_data_access_path()) {
108
0
            thrift_path.__isset.data_access_path = true;
109
0
            for (int i = 0; i < pb_path.data_access_path().path_size(); ++i) {
110
0
                thrift_path.data_access_path.path.push_back(pb_path.data_access_path().path(i));
111
0
            }
112
0
        }
113
0
        if (pb_path.has_meta_access_path()) {
114
0
            thrift_path.__isset.meta_access_path = true;
115
0
            for (int i = 0; i < pb_path.meta_access_path().path_size(); ++i) {
116
0
                thrift_path.meta_access_path.path.push_back(pb_path.meta_access_path().path(i));
117
0
            }
118
0
        }
119
0
        return thrift_path;
120
0
    };
121
84
    for (const auto& pb_path : pdesc.all_access_paths()) {
122
0
        _all_access_paths.push_back(convert_to_thrift_column_access_path(pb_path));
123
0
    }
124
84
    for (const auto& pb_path : pdesc.predicate_access_paths()) {
125
0
        _predicate_access_paths.push_back(convert_to_thrift_column_access_path(pb_path));
126
0
    }
127
84
}
128
129
#ifdef BE_TEST
130
SlotDescriptor::SlotDescriptor()
131
364
        : _id(0),
132
364
          _type(nullptr),
133
364
          _parent(0),
134
364
          _col_pos(0),
135
364
          _col_unique_id(0),
136
364
          _slot_idx(0),
137
364
          _field_idx(-1),
138
364
          _is_key(false),
139
364
          _is_auto_increment(false) {}
140
#endif
141
142
96
void SlotDescriptor::to_protobuf(PSlotDescriptor* pslot) const {
143
96
    pslot->set_id(_id);
144
96
    pslot->set_parent(_parent);
145
96
    _type->to_protobuf(pslot->mutable_slot_type());
146
96
    pslot->set_column_pos(_col_pos);
147
96
    pslot->set_byte_offset(0);
148
96
    pslot->set_null_indicator_byte(0);
149
96
    pslot->set_null_indicator_bit(_type->is_nullable() ? 0 : -1);
150
96
    pslot->set_col_name(_col_name);
151
96
    pslot->set_slot_idx(_slot_idx);
152
96
    pslot->set_col_unique_id(_col_unique_id);
153
96
    pslot->set_is_key(_is_key);
154
96
    pslot->set_is_auto_increment(_is_auto_increment);
155
96
    pslot->set_col_type(_type->get_primitive_type());
156
96
    for (const std::string& path : _column_paths) {
157
0
        pslot->add_column_paths(path);
158
0
    }
159
96
    auto convert_to_protobuf_column_access_path = [](const TColumnAccessPath& thrift_path,
160
96
                                                     doris::PColumnAccessPath* pb_path) {
161
0
        pb_path->Clear();
162
0
        pb_path->set_type((PAccessPathType)thrift_path.type); // 使用 reinterpret_cast 进行类型转换
163
0
        if (thrift_path.__isset.data_access_path) {
164
0
            auto* pb_data = pb_path->mutable_data_access_path();
165
0
            pb_data->Clear();
166
0
            for (const auto& s : thrift_path.data_access_path.path) {
167
0
                pb_data->add_path(s);
168
0
            }
169
0
        }
170
0
        if (thrift_path.__isset.meta_access_path) {
171
0
            auto* pb_meta = pb_path->mutable_meta_access_path();
172
0
            pb_meta->Clear();
173
0
            for (const auto& s : thrift_path.meta_access_path.path) {
174
0
                pb_meta->add_path(s);
175
0
            }
176
0
        }
177
0
    };
178
96
    for (const auto& path : _all_access_paths) {
179
0
        auto* pb_path = pslot->add_all_access_paths();
180
0
        convert_to_protobuf_column_access_path(path, pb_path);
181
0
    }
182
96
    for (const auto& path : _predicate_access_paths) {
183
0
        auto* pb_path = pslot->add_predicate_access_paths();
184
0
        convert_to_protobuf_column_access_path(path, pb_path);
185
0
    }
186
96
}
187
188
748k
DataTypePtr SlotDescriptor::get_data_type_ptr() const {
189
748k
    return get_data_type_with_default_argument(type());
190
748k
}
191
192
271k
MutableColumnPtr SlotDescriptor::get_empty_mutable_column() const {
193
271k
    if (this->get_virtual_column_expr() != nullptr) {
194
0
        return ColumnNothing::create(0);
195
0
    }
196
197
271k
    return type()->create_column();
198
271k
}
199
200
342
bool SlotDescriptor::is_nullable() const {
201
342
    return _type->is_nullable();
202
342
}
203
204
821
PrimitiveType SlotDescriptor::col_type() const {
205
821
    return _type->get_primitive_type();
206
821
}
207
208
8
std::string SlotDescriptor::debug_string() const {
209
8
    const bool is_virtual = this->get_virtual_column_expr() != nullptr;
210
8
    return fmt::format(
211
8
            "SlotDescriptor(id={}, type={}, col_name={}, col_unique_id={}, "
212
8
            "is_virtual={})",
213
8
            _id, _type->get_name(), _col_name, _col_unique_id, is_virtual);
214
8
}
215
216
TableDescriptor::TableDescriptor(const TTableDescriptor& tdesc)
217
218
        : _table_type(tdesc.tableType),
218
218
          _name(tdesc.tableName),
219
218
          _database(tdesc.dbName),
220
218
          _table_id(tdesc.id),
221
218
          _num_cols(tdesc.numCols),
222
218
          _num_clustering_cols(tdesc.numClusteringCols) {}
223
224
0
std::string TableDescriptor::debug_string() const {
225
0
    std::stringstream out;
226
0
    out << "#cols=" << _num_cols << " #clustering_cols=" << _num_clustering_cols;
227
0
    return out.str();
228
0
}
229
230
95
OlapTableDescriptor::OlapTableDescriptor(const TTableDescriptor& tdesc) : TableDescriptor(tdesc) {}
231
232
0
std::string OlapTableDescriptor::debug_string() const {
233
0
    std::stringstream out;
234
0
    out << "OlapTable(" << TableDescriptor::debug_string() << ")";
235
0
    return out.str();
236
0
}
237
238
DictionaryTableDescriptor::DictionaryTableDescriptor(const TTableDescriptor& tdesc)
239
0
        : TableDescriptor(tdesc) {}
240
241
0
std::string DictionaryTableDescriptor::debug_string() const {
242
0
    std::stringstream out;
243
0
    out << "Dictionary(" << TableDescriptor::debug_string() << ")";
244
0
    return out.str();
245
0
}
246
247
SchemaTableDescriptor::SchemaTableDescriptor(const TTableDescriptor& tdesc)
248
0
        : TableDescriptor(tdesc), _schema_table_type(tdesc.schemaTable.tableType) {}
249
0
SchemaTableDescriptor::~SchemaTableDescriptor() = default;
250
251
0
std::string SchemaTableDescriptor::debug_string() const {
252
0
    std::stringstream out;
253
0
    out << "SchemaTable(" << TableDescriptor::debug_string() << ")";
254
0
    return out.str();
255
0
}
256
257
BrokerTableDescriptor::BrokerTableDescriptor(const TTableDescriptor& tdesc)
258
0
        : TableDescriptor(tdesc) {}
259
260
0
BrokerTableDescriptor::~BrokerTableDescriptor() = default;
261
262
0
std::string BrokerTableDescriptor::debug_string() const {
263
0
    std::stringstream out;
264
0
    out << "BrokerTable(" << TableDescriptor::debug_string() << ")";
265
0
    return out.str();
266
0
}
267
268
36
HiveTableDescriptor::HiveTableDescriptor(const TTableDescriptor& tdesc) : TableDescriptor(tdesc) {}
269
270
36
HiveTableDescriptor::~HiveTableDescriptor() = default;
271
272
0
std::string HiveTableDescriptor::debug_string() const {
273
0
    std::stringstream out;
274
0
    out << "HiveTable(" << TableDescriptor::debug_string() << ")";
275
0
    return out.str();
276
0
}
277
278
IcebergTableDescriptor::IcebergTableDescriptor(const TTableDescriptor& tdesc)
279
0
        : TableDescriptor(tdesc) {}
280
281
0
IcebergTableDescriptor::~IcebergTableDescriptor() = default;
282
283
0
std::string IcebergTableDescriptor::debug_string() const {
284
0
    std::stringstream out;
285
0
    out << "IcebergTable(" << TableDescriptor::debug_string() << ")";
286
0
    return out.str();
287
0
}
288
289
MaxComputeTableDescriptor::MaxComputeTableDescriptor(const TTableDescriptor& tdesc)
290
0
        : TableDescriptor(tdesc),
291
0
          _region(tdesc.mcTable.region),
292
0
          _project(tdesc.mcTable.project),
293
0
          _table(tdesc.mcTable.table),
294
0
          _odps_url(tdesc.mcTable.odps_url),
295
0
          _tunnel_url(tdesc.mcTable.tunnel_url),
296
0
          _access_key(tdesc.mcTable.access_key),
297
0
          _secret_key(tdesc.mcTable.secret_key),
298
0
          _public_access(tdesc.mcTable.public_access) {
299
0
    if (tdesc.mcTable.__isset.endpoint) {
300
0
        _endpoint = tdesc.mcTable.endpoint;
301
0
    } else {
302
0
        _init_status = Status::InvalidArgument(
303
0
                "fail to init MaxComputeTableDescriptor, missing endpoint.");
304
0
    }
305
306
0
    if (tdesc.mcTable.__isset.quota) {
307
0
        _quota = tdesc.mcTable.quota;
308
0
    } else {
309
0
        _init_status =
310
0
                Status::InvalidArgument("fail to init MaxComputeTableDescriptor, missing quota.");
311
0
    }
312
313
0
    if (tdesc.mcTable.__isset.properties) [[likely]] {
314
0
        _props = tdesc.mcTable.properties;
315
0
    } else {
316
0
        static const std::string MC_ACCESS_KEY = "mc.access_key";
317
0
        static const std::string MC_SECRET_KEY = "mc.secret_key";
318
0
        _props.insert({MC_ACCESS_KEY, _access_key});
319
0
        _props.insert({MC_SECRET_KEY, _secret_key});
320
0
    }
321
0
}
322
323
0
MaxComputeTableDescriptor::~MaxComputeTableDescriptor() = default;
324
325
0
std::string MaxComputeTableDescriptor::debug_string() const {
326
0
    std::stringstream out;
327
0
    out << "MaxComputeTable(" << TableDescriptor::debug_string() << ")";
328
0
    return out.str();
329
0
}
330
331
TrinoConnectorTableDescriptor::TrinoConnectorTableDescriptor(const TTableDescriptor& tdesc)
332
0
        : TableDescriptor(tdesc) {}
333
334
0
TrinoConnectorTableDescriptor::~TrinoConnectorTableDescriptor() = default;
335
336
0
std::string TrinoConnectorTableDescriptor::debug_string() const {
337
0
    std::stringstream out;
338
0
    out << "TrinoConnectorTable(" << TableDescriptor::debug_string() << ")";
339
0
    return out.str();
340
0
}
341
342
0
EsTableDescriptor::EsTableDescriptor(const TTableDescriptor& tdesc) : TableDescriptor(tdesc) {}
343
344
0
EsTableDescriptor::~EsTableDescriptor() = default;
345
346
0
std::string EsTableDescriptor::debug_string() const {
347
0
    std::stringstream out;
348
0
    out << "EsTable(" << TableDescriptor::debug_string() << ")";
349
0
    return out.str();
350
0
}
351
352
MySQLTableDescriptor::MySQLTableDescriptor(const TTableDescriptor& tdesc)
353
87
        : TableDescriptor(tdesc),
354
87
          _mysql_db(tdesc.mysqlTable.db),
355
87
          _mysql_table(tdesc.mysqlTable.table),
356
87
          _host(tdesc.mysqlTable.host),
357
87
          _port(tdesc.mysqlTable.port),
358
87
          _user(tdesc.mysqlTable.user),
359
87
          _passwd(tdesc.mysqlTable.passwd),
360
87
          _charset(tdesc.mysqlTable.charset) {}
361
362
0
std::string MySQLTableDescriptor::debug_string() const {
363
0
    std::stringstream out;
364
0
    out << "MySQLTable(" << TableDescriptor::debug_string() << " _db" << _mysql_db
365
0
        << " table=" << _mysql_table << " host=" << _host << " port=" << _port << " user=" << _user
366
0
        << " passwd=" << _passwd << " charset=" << _charset;
367
0
    return out.str();
368
0
}
369
370
JdbcTableDescriptor::JdbcTableDescriptor(const TTableDescriptor& tdesc)
371
0
        : TableDescriptor(tdesc),
372
0
          _jdbc_catalog_id(tdesc.jdbcTable.catalog_id),
373
0
          _jdbc_resource_name(tdesc.jdbcTable.jdbc_resource_name),
374
0
          _jdbc_driver_url(tdesc.jdbcTable.jdbc_driver_url),
375
0
          _jdbc_driver_class(tdesc.jdbcTable.jdbc_driver_class),
376
0
          _jdbc_driver_checksum(tdesc.jdbcTable.jdbc_driver_checksum),
377
0
          _jdbc_url(tdesc.jdbcTable.jdbc_url),
378
0
          _jdbc_table_name(tdesc.jdbcTable.jdbc_table_name),
379
0
          _jdbc_user(tdesc.jdbcTable.jdbc_user),
380
0
          _jdbc_passwd(tdesc.jdbcTable.jdbc_password),
381
0
          _connection_pool_min_size(tdesc.jdbcTable.connection_pool_min_size),
382
0
          _connection_pool_max_size(tdesc.jdbcTable.connection_pool_max_size),
383
0
          _connection_pool_max_wait_time(tdesc.jdbcTable.connection_pool_max_wait_time),
384
0
          _connection_pool_max_life_time(tdesc.jdbcTable.connection_pool_max_life_time),
385
0
          _connection_pool_keep_alive(tdesc.jdbcTable.connection_pool_keep_alive) {}
386
387
0
std::string JdbcTableDescriptor::debug_string() const {
388
0
    fmt::memory_buffer buf;
389
0
    fmt::format_to(
390
0
            buf,
391
0
            "JDBCTable({} ,_jdbc_catalog_id = {}, _jdbc_resource_name={} ,_jdbc_driver_url={} "
392
0
            ",_jdbc_driver_class={} ,_jdbc_driver_checksum={} ,_jdbc_url={} "
393
0
            ",_jdbc_table_name={} ,_jdbc_user={} ,_jdbc_passwd={} ,_connection_pool_min_size={} "
394
0
            ",_connection_pool_max_size={} ,_connection_pool_max_wait_time={} "
395
0
            ",_connection_pool_max_life_time={} ,_connection_pool_keep_alive={})",
396
0
            TableDescriptor::debug_string(), _jdbc_catalog_id, _jdbc_resource_name,
397
0
            _jdbc_driver_url, _jdbc_driver_class, _jdbc_driver_checksum, _jdbc_url,
398
0
            _jdbc_table_name, _jdbc_user, _jdbc_passwd, _connection_pool_min_size,
399
0
            _connection_pool_max_size, _connection_pool_max_wait_time,
400
0
            _connection_pool_max_life_time, _connection_pool_keep_alive);
401
0
    return fmt::to_string(buf);
402
0
}
403
404
FilesetTableDescriptor::FilesetTableDescriptor(const TTableDescriptor& tdesc)
405
0
        : TableDescriptor(tdesc),
406
0
          _table_path(tdesc.filesetTable.table_path),
407
0
          _file_type(tdesc.filesetTable.file_type),
408
0
          _properties(tdesc.filesetTable.properties) {}
409
410
0
std::string FilesetTableDescriptor::debug_string() const {
411
0
    return fmt::format("FilesetTable({} table_path={}, file_type={})",
412
0
                       TableDescriptor::debug_string(), _table_path, static_cast<int>(_file_type));
413
0
}
414
415
RemoteDorisTableDescriptor::RemoteDorisTableDescriptor(const TTableDescriptor& tdesc)
416
0
        : TableDescriptor(tdesc) {}
417
418
0
RemoteDorisTableDescriptor::~RemoteDorisTableDescriptor() = default;
419
420
0
std::string RemoteDorisTableDescriptor::debug_string() const {
421
0
    std::stringstream out;
422
0
    out << "RemoteDorisTable(" << TableDescriptor::debug_string() << ")";
423
0
    return out.str();
424
0
}
425
426
TupleDescriptor::TupleDescriptor(const TTupleDescriptor& tdesc, bool own_slots)
427
292k
        : _id(tdesc.id),
428
292k
          _num_materialized_slots(0),
429
292k
          _has_varlen_slots(false),
430
292k
          _own_slots(own_slots) {}
431
432
TupleDescriptor::TupleDescriptor(const PTupleDescriptor& pdesc, bool own_slots)
433
18
        : _id(pdesc.id()),
434
18
          _num_materialized_slots(0),
435
18
          _has_varlen_slots(false),
436
18
          _own_slots(own_slots) {}
437
438
768k
void TupleDescriptor::add_slot(SlotDescriptor* slot) {
439
768k
    _slots.push_back(slot);
440
768k
    ++_num_materialized_slots;
441
442
768k
    if (is_complex_type(slot->type()->get_primitive_type()) ||
443
768k
        is_var_len_object(slot->type()->get_primitive_type()) ||
444
768k
        is_string_type(slot->type()->get_primitive_type())) {
445
129k
        _has_varlen_slots = true;
446
129k
    }
447
768k
}
448
449
16
void TupleDescriptor::to_protobuf(PTupleDescriptor* ptuple) const {
450
16
    ptuple->Clear();
451
16
    ptuple->set_id(_id);
452
    // Useless not set
453
16
    ptuple->set_byte_size(0);
454
16
    ptuple->set_table_id(-1);
455
16
    ptuple->set_num_null_bytes(0);
456
16
}
457
458
2
std::string TupleDescriptor::debug_string() const {
459
2
    std::stringstream out;
460
2
    out << "Tuple(id=" << _id;
461
2
    if (_table_desc != nullptr) {
462
        //out << " " << _table_desc->debug_string();
463
0
    }
464
465
2
    out << " slots=[";
466
8
    for (size_t i = 0; i < _slots.size(); ++i) {
467
6
        if (i > 0) {
468
4
            out << ", ";
469
4
        }
470
6
        out << _slots[i]->debug_string();
471
6
    }
472
473
2
    out << "]";
474
2
    out << " has_varlen_slots=" << _has_varlen_slots;
475
2
    out << ")";
476
2
    return out.str();
477
2
}
478
479
RowDescriptor::RowDescriptor(const DescriptorTbl& desc_tbl,
480
360k
                             const std::vector<TTupleId>& row_tuples) {
481
360k
    DCHECK_GT(row_tuples.size(), 0);
482
360k
    _num_materialized_slots = 0;
483
360k
    _num_slots = 0;
484
485
433k
    for (int row_tuple : row_tuples) {
486
433k
        TupleDescriptor* tupleDesc = desc_tbl.get_tuple_descriptor(row_tuple);
487
433k
        _num_materialized_slots += tupleDesc->num_materialized_slots();
488
433k
        _num_slots += tupleDesc->slots().size();
489
433k
        _tuple_desc_map.push_back(tupleDesc);
490
433k
        DCHECK(_tuple_desc_map.back() != nullptr);
491
433k
    }
492
493
360k
    init_tuple_idx_map();
494
360k
    init_has_varlen_slots();
495
360k
}
496
497
4.10k
RowDescriptor::RowDescriptor(TupleDescriptor* tuple_desc) : _tuple_desc_map(1, tuple_desc) {
498
4.10k
    init_tuple_idx_map();
499
4.10k
    init_has_varlen_slots();
500
4.10k
    _num_slots = static_cast<int32_t>(tuple_desc->slots().size());
501
4.10k
}
502
503
0
RowDescriptor::RowDescriptor(const RowDescriptor& lhs_row_desc, const RowDescriptor& rhs_row_desc) {
504
0
    _tuple_desc_map.insert(_tuple_desc_map.end(), lhs_row_desc._tuple_desc_map.begin(),
505
0
                           lhs_row_desc._tuple_desc_map.end());
506
0
    _tuple_desc_map.insert(_tuple_desc_map.end(), rhs_row_desc._tuple_desc_map.begin(),
507
0
                           rhs_row_desc._tuple_desc_map.end());
508
0
    init_tuple_idx_map();
509
0
    init_has_varlen_slots();
510
511
0
    _num_slots = lhs_row_desc.num_slots() + rhs_row_desc.num_slots();
512
0
}
513
514
365k
void RowDescriptor::init_tuple_idx_map() {
515
    // find max id
516
365k
    TupleId max_id = 0;
517
437k
    for (auto& i : _tuple_desc_map) {
518
437k
        max_id = std::max(i->id(), max_id);
519
437k
    }
520
521
365k
    _tuple_idx_map.resize(max_id + 1, INVALID_IDX);
522
802k
    for (int i = 0; i < _tuple_desc_map.size(); ++i) {
523
437k
        _tuple_idx_map[_tuple_desc_map[i]->id()] = i;
524
437k
    }
525
365k
}
526
527
365k
void RowDescriptor::init_has_varlen_slots() {
528
365k
    _has_varlen_slots = false;
529
415k
    for (auto& i : _tuple_desc_map) {
530
415k
        if (i->has_varlen_slots()) {
531
112k
            _has_varlen_slots = true;
532
112k
            break;
533
112k
        }
534
415k
    }
535
365k
}
536
537
0
int RowDescriptor::get_tuple_idx(TupleId id) const {
538
    // comment CHECK temporarily to make fuzzy test run smoothly
539
    // DCHECK_LT(id, _tuple_idx_map.size()) << "RowDescriptor: " << debug_string();
540
0
    if (_tuple_idx_map.size() <= id) {
541
0
        return RowDescriptor::INVALID_IDX;
542
0
    }
543
0
    return _tuple_idx_map[id];
544
0
}
545
546
0
void RowDescriptor::to_thrift(std::vector<TTupleId>* row_tuple_ids) {
547
0
    row_tuple_ids->clear();
548
549
0
    for (auto& i : _tuple_desc_map) {
550
0
        row_tuple_ids->push_back(i->id());
551
0
    }
552
0
}
553
554
void RowDescriptor::to_protobuf(
555
0
        google::protobuf::RepeatedField<google::protobuf::int32>* row_tuple_ids) const {
556
0
    row_tuple_ids->Clear();
557
0
    for (auto* desc : _tuple_desc_map) {
558
0
        row_tuple_ids->Add(desc->id());
559
0
    }
560
0
}
561
562
0
bool RowDescriptor::is_prefix_of(const RowDescriptor& other_desc) const {
563
0
    if (_tuple_desc_map.size() > other_desc._tuple_desc_map.size()) {
564
0
        return false;
565
0
    }
566
567
0
    for (int i = 0; i < _tuple_desc_map.size(); ++i) {
568
        // pointer comparison okay, descriptors are unique
569
0
        if (_tuple_desc_map[i] != other_desc._tuple_desc_map[i]) {
570
0
            return false;
571
0
        }
572
0
    }
573
574
0
    return true;
575
0
}
576
577
0
bool RowDescriptor::equals(const RowDescriptor& other_desc) const {
578
0
    if (_tuple_desc_map.size() != other_desc._tuple_desc_map.size()) {
579
0
        return false;
580
0
    }
581
582
0
    for (int i = 0; i < _tuple_desc_map.size(); ++i) {
583
        // pointer comparison okay, descriptors are unique
584
0
        if (_tuple_desc_map[i] != other_desc._tuple_desc_map[i]) {
585
0
            return false;
586
0
        }
587
0
    }
588
589
0
    return true;
590
0
}
591
592
0
std::string RowDescriptor::debug_string() const {
593
0
    std::stringstream ss;
594
595
0
    ss << "tuple_desc_map: [";
596
0
    for (int i = 0; i < _tuple_desc_map.size(); ++i) {
597
0
        ss << _tuple_desc_map[i]->debug_string();
598
0
        if (i != _tuple_desc_map.size() - 1) {
599
0
            ss << ", ";
600
0
        }
601
0
    }
602
0
    ss << "] ";
603
604
0
    ss << "tuple_id_map: [";
605
0
    for (int i = 0; i < _tuple_idx_map.size(); ++i) {
606
0
        ss << _tuple_idx_map[i];
607
0
        if (i != _tuple_idx_map.size() - 1) {
608
0
            ss << ", ";
609
0
        }
610
0
    }
611
0
    ss << "] ";
612
613
0
    return ss.str();
614
0
}
615
616
258k
int RowDescriptor::get_column_id(int slot_id) const {
617
258k
    int column_id_counter = 0;
618
258k
    for (auto* const tuple_desc : _tuple_desc_map) {
619
420k
        for (auto* const slot : tuple_desc->slots()) {
620
420k
            if (slot->id() == slot_id) {
621
258k
                return column_id_counter;
622
258k
            }
623
162k
            column_id_counter++;
624
162k
        }
625
258k
    }
626
96
    return -1;
627
258k
}
628
629
Status DescriptorTbl::create(ObjectPool* pool, const TDescriptorTable& thrift_tbl,
630
76.4k
                             DescriptorTbl** tbl) {
631
76.4k
    *tbl = pool->add(new DescriptorTbl());
632
633
    // deserialize table descriptors first, they are being referenced by tuple descriptors
634
76.4k
    for (const auto& tdesc : thrift_tbl.tableDescriptors) {
635
218
        TableDescriptor* desc = nullptr;
636
637
218
        switch (tdesc.tableType) {
638
87
        case TTableType::MYSQL_TABLE:
639
87
            desc = pool->add(new MySQLTableDescriptor(tdesc));
640
87
            break;
641
642
95
        case TTableType::OLAP_TABLE:
643
95
            desc = pool->add(new OlapTableDescriptor(tdesc));
644
95
            break;
645
646
0
        case TTableType::SCHEMA_TABLE:
647
0
            desc = pool->add(new SchemaTableDescriptor(tdesc));
648
0
            break;
649
0
        case TTableType::BROKER_TABLE:
650
0
            desc = pool->add(new BrokerTableDescriptor(tdesc));
651
0
            break;
652
0
        case TTableType::ES_TABLE:
653
0
            desc = pool->add(new EsTableDescriptor(tdesc));
654
0
            break;
655
36
        case TTableType::HIVE_TABLE:
656
36
            desc = pool->add(new HiveTableDescriptor(tdesc));
657
36
            break;
658
0
        case TTableType::ICEBERG_TABLE:
659
0
            desc = pool->add(new IcebergTableDescriptor(tdesc));
660
0
            break;
661
0
        case TTableType::JDBC_TABLE:
662
0
            desc = pool->add(new JdbcTableDescriptor(tdesc));
663
0
            break;
664
0
        case TTableType::FILESET_TABLE:
665
0
            desc = pool->add(new FilesetTableDescriptor(tdesc));
666
0
            break;
667
0
        case TTableType::MAX_COMPUTE_TABLE:
668
0
            desc = pool->add(new MaxComputeTableDescriptor(tdesc));
669
0
            break;
670
0
        case TTableType::TRINO_CONNECTOR_TABLE:
671
0
            desc = pool->add(new TrinoConnectorTableDescriptor(tdesc));
672
0
            break;
673
0
        case TTableType::DICTIONARY_TABLE:
674
0
            desc = pool->add(new DictionaryTableDescriptor(tdesc));
675
0
            break;
676
0
        case TTableType::REMOTE_DORIS_TABLE:
677
0
            desc = pool->add(new RemoteDorisTableDescriptor(tdesc));
678
0
            break;
679
0
        default:
680
0
            DCHECK(false) << "invalid table type: " << tdesc.tableType;
681
218
        }
682
683
218
        (*tbl)->_tbl_desc_map[static_cast<int32_t>(tdesc.id)] = desc;
684
218
    }
685
686
292k
    for (const auto& tdesc : thrift_tbl.tupleDescriptors) {
687
292k
        TupleDescriptor* desc = pool->add(new TupleDescriptor(tdesc));
688
689
        // fix up table pointer
690
292k
        if (tdesc.__isset.tableId) {
691
190
            desc->_table_desc = (*tbl)->get_table_descriptor(static_cast<int32_t>(tdesc.tableId));
692
190
            DCHECK(desc->_table_desc != nullptr);
693
190
        }
694
695
292k
        (*tbl)->_tuple_desc_map[tdesc.id] = desc;
696
292k
        (*tbl)->_row_tuples.emplace_back(tdesc.id);
697
292k
    }
698
699
768k
    for (const auto& tdesc : thrift_tbl.slotDescriptors) {
700
768k
        SlotDescriptor* slot_d = pool->add(new SlotDescriptor(tdesc));
701
768k
        (*tbl)->_slot_desc_map[tdesc.id] = slot_d;
702
703
        // link to parent
704
768k
        auto entry = (*tbl)->_tuple_desc_map.find(tdesc.parent);
705
706
768k
        if (entry == (*tbl)->_tuple_desc_map.end()) {
707
0
            return Status::InternalError("unknown tid in slot descriptor msg");
708
0
        }
709
768k
        entry->second->add_slot(slot_d);
710
768k
    }
711
712
76.4k
    return Status::OK();
713
76.4k
}
714
715
190
TableDescriptor* DescriptorTbl::get_table_descriptor(TableId id) const {
716
    // TODO: is there some boost function to do exactly this?
717
190
    auto i = _tbl_desc_map.find(id);
718
719
190
    if (i == _tbl_desc_map.end()) {
720
0
        return nullptr;
721
190
    } else {
722
190
        return i->second;
723
190
    }
724
190
}
725
726
639k
TupleDescriptor* DescriptorTbl::get_tuple_descriptor(TupleId id) const {
727
    // TODO: is there some boost function to do exactly this?
728
639k
    auto i = _tuple_desc_map.find(id);
729
730
639k
    if (i == _tuple_desc_map.end()) {
731
3
        return nullptr;
732
639k
    } else {
733
639k
        return i->second;
734
639k
    }
735
639k
}
736
737
258k
SlotDescriptor* DescriptorTbl::get_slot_descriptor(SlotId id) const {
738
    // TODO: is there some boost function to do exactly this?
739
258k
    auto i = _slot_desc_map.find(id);
740
741
258k
    if (i == _slot_desc_map.end()) {
742
0
        return nullptr;
743
258k
    } else {
744
258k
        return i->second;
745
258k
    }
746
258k
}
747
748
0
std::string DescriptorTbl::debug_string() const {
749
0
    std::stringstream out;
750
0
    out << "tuples:\n";
751
752
0
    for (auto i : _tuple_desc_map) {
753
0
        out << i.second->debug_string() << '\n';
754
0
    }
755
756
0
    return out.str();
757
0
}
758
} // namespace doris