Coverage Report

Created: 2025-12-03 20:21

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/root/doris/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 "util/string_util.h"
36
#include "vec/aggregate_functions/aggregate_function.h"
37
#include "vec/columns/column_nothing.h"
38
#include "vec/core/types.h"
39
#include "vec/data_types/data_type_array.h"
40
#include "vec/data_types/data_type_decimal.h"
41
#include "vec/data_types/data_type_factory.hpp"
42
#include "vec/data_types/data_type_map.h"
43
#include "vec/data_types/data_type_struct.h"
44
#include "vec/exprs/vexpr.h"
45
#include "vec/functions/function_helpers.h"
46
#include "vec/utils/util.hpp"
47
48
namespace doris {
49
#include "common/compile_check_begin.h"
50
const int RowDescriptor::INVALID_IDX = -1;
51
52
SlotDescriptor::SlotDescriptor(const TSlotDescriptor& tdesc)
53
768k
        : _id(tdesc.id),
54
768k
          _type(vectorized::DataTypeFactory::instance().create_data_type(
55
768k
                  tdesc.slotType, tdesc.nullIndicatorBit != -1)),
56
768k
          _parent(tdesc.parent),
57
768k
          _col_pos(tdesc.columnPos),
58
768k
          _col_name(tdesc.colName),
59
768k
          _col_name_lower_case(to_lower(tdesc.colName)),
60
768k
          _col_unique_id(tdesc.col_unique_id),
61
768k
          _slot_idx(tdesc.slotIdx),
62
768k
          _field_idx(-1),
63
768k
          _is_key(tdesc.is_key),
64
768k
          _column_paths(tdesc.column_paths),
65
768k
          _all_access_paths(tdesc.__isset.all_access_paths ? tdesc.all_access_paths
66
768k
                                                           : TColumnAccessPaths {}),
67
768k
          _predicate_access_paths(tdesc.__isset.predicate_access_paths
68
768k
                                          ? tdesc.predicate_access_paths
69
768k
                                          : TColumnAccessPaths {}),
70
768k
          _is_auto_increment(tdesc.__isset.is_auto_increment ? tdesc.is_auto_increment : false),
71
768k
          _col_default_value(tdesc.__isset.col_default_value ? tdesc.col_default_value : "") {
72
768k
    if (tdesc.__isset.virtual_column_expr) {
73
        // Make sure virtual column is valid.
74
97
        if (tdesc.virtual_column_expr.nodes.empty()) {
75
2
            throw doris::Exception(doris::ErrorCode::FATAL_ERROR,
76
2
                                   "Virtual column expr node is empty, col_name: {}, "
77
2
                                   "col_unique_id: {}",
78
2
                                   tdesc.colName, tdesc.col_unique_id);
79
2
        }
80
95
        const auto& node = tdesc.virtual_column_expr.nodes[0];
81
95
        if (node.node_type == TExprNodeType::SLOT_REF) {
82
2
            throw doris::Exception(doris::ErrorCode::FATAL_ERROR,
83
2
                                   "Virtual column expr node is slot ref, col_name: {}, "
84
2
                                   "col_unique_id: {}",
85
2
                                   tdesc.colName, tdesc.col_unique_id);
86
2
        }
87
93
        this->virtual_column_expr = std::make_shared<doris::TExpr>(tdesc.virtual_column_expr);
88
93
    }
89
768k
}
90
91
SlotDescriptor::SlotDescriptor(const PSlotDescriptor& pdesc)
92
84
        : _id(pdesc.id()),
93
84
          _type(vectorized::DataTypeFactory::instance().create_data_type(
94
84
                  pdesc.slot_type(), pdesc.null_indicator_bit() != -1)),
95
84
          _parent(pdesc.parent()),
96
84
          _col_pos(pdesc.column_pos()),
97
84
          _col_name(pdesc.col_name()),
98
84
          _col_name_lower_case(to_lower(pdesc.col_name())),
99
84
          _col_unique_id(pdesc.col_unique_id()),
100
84
          _slot_idx(pdesc.slot_idx()),
101
84
          _field_idx(-1),
102
84
          _is_key(pdesc.is_key()),
103
84
          _column_paths(pdesc.column_paths().begin(), pdesc.column_paths().end()),
104
84
          _is_auto_increment(pdesc.is_auto_increment()) {
105
84
    auto convert_to_thrift_column_access_path = [](const PColumnAccessPath& pb_path) {
106
0
        TColumnAccessPath thrift_path;
107
0
        thrift_path.type = (TAccessPathType::type)pb_path.type();
108
0
        if (pb_path.has_data_access_path()) {
109
0
            thrift_path.__isset.data_access_path = true;
110
0
            for (int i = 0; i < pb_path.data_access_path().path_size(); ++i) {
111
0
                thrift_path.data_access_path.path.push_back(pb_path.data_access_path().path(i));
112
0
            }
113
0
        }
114
0
        if (pb_path.has_meta_access_path()) {
115
0
            thrift_path.__isset.meta_access_path = true;
116
0
            for (int i = 0; i < pb_path.meta_access_path().path_size(); ++i) {
117
0
                thrift_path.meta_access_path.path.push_back(pb_path.meta_access_path().path(i));
118
0
            }
119
0
        }
120
0
        return thrift_path;
121
0
    };
122
84
    for (const auto& pb_path : pdesc.all_access_paths()) {
123
0
        _all_access_paths.push_back(convert_to_thrift_column_access_path(pb_path));
124
0
    }
125
84
    for (const auto& pb_path : pdesc.predicate_access_paths()) {
126
0
        _predicate_access_paths.push_back(convert_to_thrift_column_access_path(pb_path));
127
0
    }
128
84
}
129
130
#ifdef BE_TEST
131
SlotDescriptor::SlotDescriptor()
132
299
        : _id(0),
133
299
          _type(nullptr),
134
299
          _parent(0),
135
299
          _col_pos(0),
136
299
          _col_unique_id(0),
137
299
          _slot_idx(0),
138
299
          _field_idx(-1),
139
299
          _is_key(false),
140
299
          _is_auto_increment(false) {}
141
#endif
142
143
96
void SlotDescriptor::to_protobuf(PSlotDescriptor* pslot) const {
144
96
    pslot->set_id(_id);
145
96
    pslot->set_parent(_parent);
146
96
    _type->to_protobuf(pslot->mutable_slot_type());
147
96
    pslot->set_column_pos(_col_pos);
148
96
    pslot->set_byte_offset(0);
149
96
    pslot->set_null_indicator_byte(0);
150
96
    pslot->set_null_indicator_bit(_type->is_nullable() ? 0 : -1);
151
96
    pslot->set_col_name(_col_name);
152
96
    pslot->set_slot_idx(_slot_idx);
153
96
    pslot->set_col_unique_id(_col_unique_id);
154
96
    pslot->set_is_key(_is_key);
155
96
    pslot->set_is_auto_increment(_is_auto_increment);
156
96
    pslot->set_col_type(_type->get_primitive_type());
157
96
    for (const std::string& path : _column_paths) {
158
0
        pslot->add_column_paths(path);
159
0
    }
160
96
    auto convert_to_protobuf_column_access_path = [](const TColumnAccessPath& thrift_path,
161
96
                                                     doris::PColumnAccessPath* pb_path) {
162
0
        pb_path->Clear();
163
0
        pb_path->set_type((PAccessPathType)thrift_path.type); // 使用 reinterpret_cast 进行类型转换
164
0
        if (thrift_path.__isset.data_access_path) {
165
0
            auto* pb_data = pb_path->mutable_data_access_path();
166
0
            pb_data->Clear();
167
0
            for (const auto& s : thrift_path.data_access_path.path) {
168
0
                pb_data->add_path(s);
169
0
            }
170
0
        }
171
0
        if (thrift_path.__isset.meta_access_path) {
172
0
            auto* pb_meta = pb_path->mutable_meta_access_path();
173
0
            pb_meta->Clear();
174
0
            for (const auto& s : thrift_path.meta_access_path.path) {
175
0
                pb_meta->add_path(s);
176
0
            }
177
0
        }
178
0
    };
179
96
    for (const auto& path : _all_access_paths) {
180
0
        auto* pb_path = pslot->add_all_access_paths();
181
0
        convert_to_protobuf_column_access_path(path, pb_path);
182
0
    }
183
96
    for (const auto& path : _predicate_access_paths) {
184
0
        auto* pb_path = pslot->add_predicate_access_paths();
185
0
        convert_to_protobuf_column_access_path(path, pb_path);
186
0
    }
187
96
}
188
189
748k
vectorized::DataTypePtr SlotDescriptor::get_data_type_ptr() const {
190
748k
    return vectorized::get_data_type_with_default_argument(type());
191
748k
}
192
193
271k
vectorized::MutableColumnPtr SlotDescriptor::get_empty_mutable_column() const {
194
271k
    if (this->get_virtual_column_expr() != nullptr) {
195
0
        return vectorized::ColumnNothing::create(0);
196
0
    }
197
198
271k
    return type()->create_column();
199
271k
}
200
201
363
bool SlotDescriptor::is_nullable() const {
202
363
    return _type->is_nullable();
203
363
}
204
205
793
PrimitiveType SlotDescriptor::col_type() const {
206
793
    return _type->get_primitive_type();
207
793
}
208
209
8
std::string SlotDescriptor::debug_string() const {
210
8
    const bool is_virtual = this->get_virtual_column_expr() != nullptr;
211
8
    return fmt::format(
212
8
            "SlotDescriptor(id={}, type={}, col_name={}, col_unique_id={}, "
213
8
            "is_virtual={})",
214
8
            _id, _type->get_name(), _col_name, _col_unique_id, is_virtual);
215
8
}
216
217
TableDescriptor::TableDescriptor(const TTableDescriptor& tdesc)
218
193
        : _table_type(tdesc.tableType),
219
193
          _name(tdesc.tableName),
220
193
          _database(tdesc.dbName),
221
193
          _table_id(tdesc.id),
222
193
          _num_cols(tdesc.numCols),
223
193
          _num_clustering_cols(tdesc.numClusteringCols) {}
224
225
0
std::string TableDescriptor::debug_string() const {
226
0
    std::stringstream out;
227
0
    out << "#cols=" << _num_cols << " #clustering_cols=" << _num_clustering_cols;
228
0
    return out.str();
229
0
}
230
231
73
OlapTableDescriptor::OlapTableDescriptor(const TTableDescriptor& tdesc) : TableDescriptor(tdesc) {}
232
233
0
std::string OlapTableDescriptor::debug_string() const {
234
0
    std::stringstream out;
235
0
    out << "OlapTable(" << TableDescriptor::debug_string() << ")";
236
0
    return out.str();
237
0
}
238
239
DictionaryTableDescriptor::DictionaryTableDescriptor(const TTableDescriptor& tdesc)
240
0
        : TableDescriptor(tdesc) {}
241
242
0
std::string DictionaryTableDescriptor::debug_string() const {
243
0
    std::stringstream out;
244
0
    out << "Dictionary(" << TableDescriptor::debug_string() << ")";
245
0
    return out.str();
246
0
}
247
248
SchemaTableDescriptor::SchemaTableDescriptor(const TTableDescriptor& tdesc)
249
0
        : TableDescriptor(tdesc), _schema_table_type(tdesc.schemaTable.tableType) {}
250
0
SchemaTableDescriptor::~SchemaTableDescriptor() = default;
251
252
0
std::string SchemaTableDescriptor::debug_string() const {
253
0
    std::stringstream out;
254
0
    out << "SchemaTable(" << TableDescriptor::debug_string() << ")";
255
0
    return out.str();
256
0
}
257
258
BrokerTableDescriptor::BrokerTableDescriptor(const TTableDescriptor& tdesc)
259
0
        : TableDescriptor(tdesc) {}
260
261
0
BrokerTableDescriptor::~BrokerTableDescriptor() = default;
262
263
0
std::string BrokerTableDescriptor::debug_string() const {
264
0
    std::stringstream out;
265
0
    out << "BrokerTable(" << TableDescriptor::debug_string() << ")";
266
0
    return out.str();
267
0
}
268
269
36
HiveTableDescriptor::HiveTableDescriptor(const TTableDescriptor& tdesc) : TableDescriptor(tdesc) {}
270
271
36
HiveTableDescriptor::~HiveTableDescriptor() = default;
272
273
0
std::string HiveTableDescriptor::debug_string() const {
274
0
    std::stringstream out;
275
0
    out << "HiveTable(" << TableDescriptor::debug_string() << ")";
276
0
    return out.str();
277
0
}
278
279
IcebergTableDescriptor::IcebergTableDescriptor(const TTableDescriptor& tdesc)
280
0
        : TableDescriptor(tdesc) {}
281
282
0
IcebergTableDescriptor::~IcebergTableDescriptor() = default;
283
284
0
std::string IcebergTableDescriptor::debug_string() const {
285
0
    std::stringstream out;
286
0
    out << "IcebergTable(" << TableDescriptor::debug_string() << ")";
287
0
    return out.str();
288
0
}
289
290
MaxComputeTableDescriptor::MaxComputeTableDescriptor(const TTableDescriptor& tdesc)
291
0
        : TableDescriptor(tdesc),
292
0
          _region(tdesc.mcTable.region),
293
0
          _project(tdesc.mcTable.project),
294
0
          _table(tdesc.mcTable.table),
295
0
          _odps_url(tdesc.mcTable.odps_url),
296
0
          _tunnel_url(tdesc.mcTable.tunnel_url),
297
0
          _access_key(tdesc.mcTable.access_key),
298
0
          _secret_key(tdesc.mcTable.secret_key),
299
0
          _public_access(tdesc.mcTable.public_access) {
300
0
    if (tdesc.mcTable.__isset.endpoint) {
301
0
        _endpoint = tdesc.mcTable.endpoint;
302
0
    } else {
303
0
        _init_status = Status::InvalidArgument(
304
0
                "fail to init MaxComputeTableDescriptor, missing endpoint.");
305
0
    }
306
307
0
    if (tdesc.mcTable.__isset.quota) {
308
0
        _quota = tdesc.mcTable.quota;
309
0
    } else {
310
0
        _init_status =
311
0
                Status::InvalidArgument("fail to init MaxComputeTableDescriptor, missing quota.");
312
0
    }
313
0
}
314
315
0
MaxComputeTableDescriptor::~MaxComputeTableDescriptor() = default;
316
317
0
std::string MaxComputeTableDescriptor::debug_string() const {
318
0
    std::stringstream out;
319
0
    out << "MaxComputeTable(" << TableDescriptor::debug_string() << ")";
320
0
    return out.str();
321
0
}
322
323
TrinoConnectorTableDescriptor::TrinoConnectorTableDescriptor(const TTableDescriptor& tdesc)
324
0
        : TableDescriptor(tdesc) {}
325
326
0
TrinoConnectorTableDescriptor::~TrinoConnectorTableDescriptor() = default;
327
328
0
std::string TrinoConnectorTableDescriptor::debug_string() const {
329
0
    std::stringstream out;
330
0
    out << "TrinoConnectorTable(" << TableDescriptor::debug_string() << ")";
331
0
    return out.str();
332
0
}
333
334
0
EsTableDescriptor::EsTableDescriptor(const TTableDescriptor& tdesc) : TableDescriptor(tdesc) {}
335
336
0
EsTableDescriptor::~EsTableDescriptor() = default;
337
338
0
std::string EsTableDescriptor::debug_string() const {
339
0
    std::stringstream out;
340
0
    out << "EsTable(" << TableDescriptor::debug_string() << ")";
341
0
    return out.str();
342
0
}
343
344
MySQLTableDescriptor::MySQLTableDescriptor(const TTableDescriptor& tdesc)
345
84
        : TableDescriptor(tdesc),
346
84
          _mysql_db(tdesc.mysqlTable.db),
347
84
          _mysql_table(tdesc.mysqlTable.table),
348
84
          _host(tdesc.mysqlTable.host),
349
84
          _port(tdesc.mysqlTable.port),
350
84
          _user(tdesc.mysqlTable.user),
351
84
          _passwd(tdesc.mysqlTable.passwd),
352
84
          _charset(tdesc.mysqlTable.charset) {}
353
354
0
std::string MySQLTableDescriptor::debug_string() const {
355
0
    std::stringstream out;
356
0
    out << "MySQLTable(" << TableDescriptor::debug_string() << " _db" << _mysql_db
357
0
        << " table=" << _mysql_table << " host=" << _host << " port=" << _port << " user=" << _user
358
0
        << " passwd=" << _passwd << " charset=" << _charset;
359
0
    return out.str();
360
0
}
361
362
JdbcTableDescriptor::JdbcTableDescriptor(const TTableDescriptor& tdesc)
363
0
        : TableDescriptor(tdesc),
364
0
          _jdbc_catalog_id(tdesc.jdbcTable.catalog_id),
365
0
          _jdbc_resource_name(tdesc.jdbcTable.jdbc_resource_name),
366
0
          _jdbc_driver_url(tdesc.jdbcTable.jdbc_driver_url),
367
0
          _jdbc_driver_class(tdesc.jdbcTable.jdbc_driver_class),
368
0
          _jdbc_driver_checksum(tdesc.jdbcTable.jdbc_driver_checksum),
369
0
          _jdbc_url(tdesc.jdbcTable.jdbc_url),
370
0
          _jdbc_table_name(tdesc.jdbcTable.jdbc_table_name),
371
0
          _jdbc_user(tdesc.jdbcTable.jdbc_user),
372
0
          _jdbc_passwd(tdesc.jdbcTable.jdbc_password),
373
0
          _connection_pool_min_size(tdesc.jdbcTable.connection_pool_min_size),
374
0
          _connection_pool_max_size(tdesc.jdbcTable.connection_pool_max_size),
375
0
          _connection_pool_max_wait_time(tdesc.jdbcTable.connection_pool_max_wait_time),
376
0
          _connection_pool_max_life_time(tdesc.jdbcTable.connection_pool_max_life_time),
377
0
          _connection_pool_keep_alive(tdesc.jdbcTable.connection_pool_keep_alive) {}
378
379
0
std::string JdbcTableDescriptor::debug_string() const {
380
0
    fmt::memory_buffer buf;
381
0
    fmt::format_to(
382
0
            buf,
383
0
            "JDBCTable({} ,_jdbc_catalog_id = {}, _jdbc_resource_name={} ,_jdbc_driver_url={} "
384
0
            ",_jdbc_driver_class={} ,_jdbc_driver_checksum={} ,_jdbc_url={} "
385
0
            ",_jdbc_table_name={} ,_jdbc_user={} ,_jdbc_passwd={} ,_connection_pool_min_size={} "
386
0
            ",_connection_pool_max_size={} ,_connection_pool_max_wait_time={} "
387
0
            ",_connection_pool_max_life_time={} ,_connection_pool_keep_alive={})",
388
0
            TableDescriptor::debug_string(), _jdbc_catalog_id, _jdbc_resource_name,
389
0
            _jdbc_driver_url, _jdbc_driver_class, _jdbc_driver_checksum, _jdbc_url,
390
0
            _jdbc_table_name, _jdbc_user, _jdbc_passwd, _connection_pool_min_size,
391
0
            _connection_pool_max_size, _connection_pool_max_wait_time,
392
0
            _connection_pool_max_life_time, _connection_pool_keep_alive);
393
0
    return fmt::to_string(buf);
394
0
}
395
396
RemoteDorisTableDescriptor::RemoteDorisTableDescriptor(const TTableDescriptor& tdesc)
397
0
        : TableDescriptor(tdesc) {}
398
399
0
RemoteDorisTableDescriptor::~RemoteDorisTableDescriptor() = default;
400
401
0
std::string RemoteDorisTableDescriptor::debug_string() const {
402
0
    std::stringstream out;
403
0
    out << "RemoteDorisTable(" << TableDescriptor::debug_string() << ")";
404
0
    return out.str();
405
0
}
406
407
TupleDescriptor::TupleDescriptor(const TTupleDescriptor& tdesc, bool own_slots)
408
292k
        : _id(tdesc.id),
409
292k
          _num_materialized_slots(0),
410
292k
          _has_varlen_slots(false),
411
292k
          _own_slots(own_slots) {}
412
413
TupleDescriptor::TupleDescriptor(const PTupleDescriptor& pdesc, bool own_slots)
414
18
        : _id(pdesc.id()),
415
18
          _num_materialized_slots(0),
416
18
          _has_varlen_slots(false),
417
18
          _own_slots(own_slots) {}
418
419
768k
void TupleDescriptor::add_slot(SlotDescriptor* slot) {
420
768k
    _slots.push_back(slot);
421
768k
    ++_num_materialized_slots;
422
423
768k
    if (is_complex_type(slot->type()->get_primitive_type()) ||
424
768k
        is_var_len_object(slot->type()->get_primitive_type()) ||
425
768k
        is_string_type(slot->type()->get_primitive_type())) {
426
129k
        _has_varlen_slots = true;
427
129k
    }
428
768k
}
429
430
16
void TupleDescriptor::to_protobuf(PTupleDescriptor* ptuple) const {
431
16
    ptuple->Clear();
432
16
    ptuple->set_id(_id);
433
    // Useless not set
434
16
    ptuple->set_byte_size(0);
435
16
    ptuple->set_table_id(-1);
436
16
    ptuple->set_num_null_bytes(0);
437
16
}
438
439
2
std::string TupleDescriptor::debug_string() const {
440
2
    std::stringstream out;
441
2
    out << "Tuple(id=" << _id;
442
2
    if (_table_desc != nullptr) {
443
        //out << " " << _table_desc->debug_string();
444
0
    }
445
446
2
    out << " slots=[";
447
8
    for (size_t i = 0; i < _slots.size(); ++i) {
448
6
        if (i > 0) {
449
4
            out << ", ";
450
4
        }
451
6
        out << _slots[i]->debug_string();
452
6
    }
453
454
2
    out << "]";
455
2
    out << " has_varlen_slots=" << _has_varlen_slots;
456
2
    out << ")";
457
2
    return out.str();
458
2
}
459
460
RowDescriptor::RowDescriptor(const DescriptorTbl& desc_tbl, const std::vector<TTupleId>& row_tuples,
461
                             const std::vector<bool>& nullable_tuples)
462
432k
        : _tuple_idx_nullable_map(nullable_tuples) {
463
432k
    DCHECK(nullable_tuples.size() == row_tuples.size())
464
0
            << "nullable_tuples size " << nullable_tuples.size() << " != row_tuples size "
465
0
            << row_tuples.size();
466
432k
    DCHECK_GT(row_tuples.size(), 0);
467
432k
    _num_materialized_slots = 0;
468
432k
    _num_slots = 0;
469
470
504k
    for (int row_tuple : row_tuples) {
471
504k
        TupleDescriptor* tupleDesc = desc_tbl.get_tuple_descriptor(row_tuple);
472
504k
        _num_materialized_slots += tupleDesc->num_materialized_slots();
473
504k
        _num_slots += tupleDesc->slots().size();
474
504k
        _tuple_desc_map.push_back(tupleDesc);
475
504k
        DCHECK(_tuple_desc_map.back() != nullptr);
476
504k
    }
477
478
432k
    init_tuple_idx_map();
479
432k
    init_has_varlen_slots();
480
432k
}
481
482
RowDescriptor::RowDescriptor(TupleDescriptor* tuple_desc, bool is_nullable)
483
4.09k
        : _tuple_desc_map(1, tuple_desc), _tuple_idx_nullable_map(1, is_nullable) {
484
4.09k
    init_tuple_idx_map();
485
4.09k
    init_has_varlen_slots();
486
4.09k
    _num_slots = static_cast<int32_t>(tuple_desc->slots().size());
487
4.09k
}
488
489
0
RowDescriptor::RowDescriptor(const RowDescriptor& lhs_row_desc, const RowDescriptor& rhs_row_desc) {
490
0
    _tuple_desc_map.insert(_tuple_desc_map.end(), lhs_row_desc._tuple_desc_map.begin(),
491
0
                           lhs_row_desc._tuple_desc_map.end());
492
0
    _tuple_desc_map.insert(_tuple_desc_map.end(), rhs_row_desc._tuple_desc_map.begin(),
493
0
                           rhs_row_desc._tuple_desc_map.end());
494
0
    _tuple_idx_nullable_map.insert(_tuple_idx_nullable_map.end(),
495
0
                                   lhs_row_desc._tuple_idx_nullable_map.begin(),
496
0
                                   lhs_row_desc._tuple_idx_nullable_map.end());
497
0
    _tuple_idx_nullable_map.insert(_tuple_idx_nullable_map.end(),
498
0
                                   rhs_row_desc._tuple_idx_nullable_map.begin(),
499
0
                                   rhs_row_desc._tuple_idx_nullable_map.end());
500
0
    init_tuple_idx_map();
501
0
    init_has_varlen_slots();
502
503
0
    _num_slots = lhs_row_desc.num_slots() + rhs_row_desc.num_slots();
504
0
}
505
506
436k
void RowDescriptor::init_tuple_idx_map() {
507
    // find max id
508
436k
    TupleId max_id = 0;
509
508k
    for (auto& i : _tuple_desc_map) {
510
508k
        max_id = std::max(i->id(), max_id);
511
508k
    }
512
513
436k
    _tuple_idx_map.resize(max_id + 1, INVALID_IDX);
514
945k
    for (int i = 0; i < _tuple_desc_map.size(); ++i) {
515
508k
        _tuple_idx_map[_tuple_desc_map[i]->id()] = i;
516
508k
    }
517
436k
}
518
519
436k
void RowDescriptor::init_has_varlen_slots() {
520
436k
    _has_varlen_slots = false;
521
487k
    for (auto& i : _tuple_desc_map) {
522
487k
        if (i->has_varlen_slots()) {
523
133k
            _has_varlen_slots = true;
524
133k
            break;
525
133k
        }
526
487k
    }
527
436k
}
528
529
0
int RowDescriptor::get_tuple_idx(TupleId id) const {
530
    // comment CHECK temporarily to make fuzzy test run smoothly
531
    // DCHECK_LT(id, _tuple_idx_map.size()) << "RowDescriptor: " << debug_string();
532
0
    if (_tuple_idx_map.size() <= id) {
533
0
        return RowDescriptor::INVALID_IDX;
534
0
    }
535
0
    return _tuple_idx_map[id];
536
0
}
537
538
0
void RowDescriptor::to_thrift(std::vector<TTupleId>* row_tuple_ids) {
539
0
    row_tuple_ids->clear();
540
541
0
    for (auto& i : _tuple_desc_map) {
542
0
        row_tuple_ids->push_back(i->id());
543
0
    }
544
0
}
545
546
void RowDescriptor::to_protobuf(
547
0
        google::protobuf::RepeatedField<google::protobuf::int32>* row_tuple_ids) const {
548
0
    row_tuple_ids->Clear();
549
0
    for (auto* desc : _tuple_desc_map) {
550
0
        row_tuple_ids->Add(desc->id());
551
0
    }
552
0
}
553
554
0
bool RowDescriptor::is_prefix_of(const RowDescriptor& other_desc) const {
555
0
    if (_tuple_desc_map.size() > other_desc._tuple_desc_map.size()) {
556
0
        return false;
557
0
    }
558
559
0
    for (int i = 0; i < _tuple_desc_map.size(); ++i) {
560
        // pointer comparison okay, descriptors are unique
561
0
        if (_tuple_desc_map[i] != other_desc._tuple_desc_map[i]) {
562
0
            return false;
563
0
        }
564
0
    }
565
566
0
    return true;
567
0
}
568
569
0
bool RowDescriptor::equals(const RowDescriptor& other_desc) const {
570
0
    if (_tuple_desc_map.size() != other_desc._tuple_desc_map.size()) {
571
0
        return false;
572
0
    }
573
574
0
    for (int i = 0; i < _tuple_desc_map.size(); ++i) {
575
        // pointer comparison okay, descriptors are unique
576
0
        if (_tuple_desc_map[i] != other_desc._tuple_desc_map[i]) {
577
0
            return false;
578
0
        }
579
0
    }
580
581
0
    return true;
582
0
}
583
584
0
std::string RowDescriptor::debug_string() const {
585
0
    std::stringstream ss;
586
587
0
    ss << "tuple_desc_map: [";
588
0
    for (int i = 0; i < _tuple_desc_map.size(); ++i) {
589
0
        ss << _tuple_desc_map[i]->debug_string();
590
0
        if (i != _tuple_desc_map.size() - 1) {
591
0
            ss << ", ";
592
0
        }
593
0
    }
594
0
    ss << "] ";
595
596
0
    ss << "tuple_id_map: [";
597
0
    for (int i = 0; i < _tuple_idx_map.size(); ++i) {
598
0
        ss << _tuple_idx_map[i];
599
0
        if (i != _tuple_idx_map.size() - 1) {
600
0
            ss << ", ";
601
0
        }
602
0
    }
603
0
    ss << "] ";
604
605
0
    ss << "tuple_is_nullable: [";
606
0
    for (int i = 0; i < _tuple_idx_nullable_map.size(); ++i) {
607
0
        ss << _tuple_idx_nullable_map[i];
608
0
        if (i != _tuple_idx_nullable_map.size() - 1) {
609
0
            ss << ", ";
610
0
        }
611
0
    }
612
0
    ss << "] ";
613
614
0
    return ss.str();
615
0
}
616
617
258k
int RowDescriptor::get_column_id(int slot_id) const {
618
258k
    int column_id_counter = 0;
619
258k
    for (auto* const tuple_desc : _tuple_desc_map) {
620
420k
        for (auto* const slot : tuple_desc->slots()) {
621
420k
            if (slot->id() == slot_id) {
622
258k
                return column_id_counter;
623
258k
            }
624
162k
            column_id_counter++;
625
162k
        }
626
258k
    }
627
0
    return -1;
628
258k
}
629
630
Status DescriptorTbl::create(ObjectPool* pool, const TDescriptorTable& thrift_tbl,
631
76.3k
                             DescriptorTbl** tbl) {
632
76.3k
    *tbl = pool->add(new DescriptorTbl());
633
634
    // deserialize table descriptors first, they are being referenced by tuple descriptors
635
76.3k
    for (const auto& tdesc : thrift_tbl.tableDescriptors) {
636
193
        TableDescriptor* desc = nullptr;
637
638
193
        switch (tdesc.tableType) {
639
84
        case TTableType::MYSQL_TABLE:
640
84
            desc = pool->add(new MySQLTableDescriptor(tdesc));
641
84
            break;
642
643
73
        case TTableType::OLAP_TABLE:
644
73
            desc = pool->add(new OlapTableDescriptor(tdesc));
645
73
            break;
646
647
0
        case TTableType::SCHEMA_TABLE:
648
0
            desc = pool->add(new SchemaTableDescriptor(tdesc));
649
0
            break;
650
0
        case TTableType::BROKER_TABLE:
651
0
            desc = pool->add(new BrokerTableDescriptor(tdesc));
652
0
            break;
653
0
        case TTableType::ES_TABLE:
654
0
            desc = pool->add(new EsTableDescriptor(tdesc));
655
0
            break;
656
36
        case TTableType::HIVE_TABLE:
657
36
            desc = pool->add(new HiveTableDescriptor(tdesc));
658
36
            break;
659
0
        case TTableType::ICEBERG_TABLE:
660
0
            desc = pool->add(new IcebergTableDescriptor(tdesc));
661
0
            break;
662
0
        case TTableType::JDBC_TABLE:
663
0
            desc = pool->add(new JdbcTableDescriptor(tdesc));
664
0
            break;
665
0
        case TTableType::MAX_COMPUTE_TABLE:
666
0
            desc = pool->add(new MaxComputeTableDescriptor(tdesc));
667
0
            break;
668
0
        case TTableType::TRINO_CONNECTOR_TABLE:
669
0
            desc = pool->add(new TrinoConnectorTableDescriptor(tdesc));
670
0
            break;
671
0
        case TTableType::DICTIONARY_TABLE:
672
0
            desc = pool->add(new DictionaryTableDescriptor(tdesc));
673
0
            break;
674
0
        case TTableType::REMOTE_DORIS_TABLE:
675
0
            desc = pool->add(new RemoteDorisTableDescriptor(tdesc));
676
0
            break;
677
0
        default:
678
0
            DCHECK(false) << "invalid table type: " << tdesc.tableType;
679
193
        }
680
681
193
        (*tbl)->_tbl_desc_map[static_cast<int32_t>(tdesc.id)] = desc;
682
193
    }
683
684
292k
    for (const auto& tdesc : thrift_tbl.tupleDescriptors) {
685
292k
        TupleDescriptor* desc = pool->add(new TupleDescriptor(tdesc));
686
687
        // fix up table pointer
688
292k
        if (tdesc.__isset.tableId) {
689
183
            desc->_table_desc = (*tbl)->get_table_descriptor(static_cast<int32_t>(tdesc.tableId));
690
183
            DCHECK(desc->_table_desc != nullptr);
691
183
        }
692
693
292k
        (*tbl)->_tuple_desc_map[tdesc.id] = desc;
694
292k
        (*tbl)->_row_tuples.emplace_back(tdesc.id);
695
292k
    }
696
697
767k
    for (const auto& tdesc : thrift_tbl.slotDescriptors) {
698
767k
        SlotDescriptor* slot_d = pool->add(new SlotDescriptor(tdesc));
699
767k
        (*tbl)->_slot_desc_map[tdesc.id] = slot_d;
700
701
        // link to parent
702
767k
        auto entry = (*tbl)->_tuple_desc_map.find(tdesc.parent);
703
704
767k
        if (entry == (*tbl)->_tuple_desc_map.end()) {
705
0
            return Status::InternalError("unknown tid in slot descriptor msg");
706
0
        }
707
767k
        entry->second->add_slot(slot_d);
708
767k
    }
709
710
76.3k
    return Status::OK();
711
76.3k
}
712
713
183
TableDescriptor* DescriptorTbl::get_table_descriptor(TableId id) const {
714
    // TODO: is there some boost function to do exactly this?
715
183
    auto i = _tbl_desc_map.find(id);
716
717
183
    if (i == _tbl_desc_map.end()) {
718
0
        return nullptr;
719
183
    } else {
720
183
        return i->second;
721
183
    }
722
183
}
723
724
711k
TupleDescriptor* DescriptorTbl::get_tuple_descriptor(TupleId id) const {
725
    // TODO: is there some boost function to do exactly this?
726
711k
    auto i = _tuple_desc_map.find(id);
727
728
711k
    if (i == _tuple_desc_map.end()) {
729
4
        return nullptr;
730
711k
    } else {
731
711k
        return i->second;
732
711k
    }
733
711k
}
734
735
258k
SlotDescriptor* DescriptorTbl::get_slot_descriptor(SlotId id) const {
736
    // TODO: is there some boost function to do exactly this?
737
258k
    auto i = _slot_desc_map.find(id);
738
739
258k
    if (i == _slot_desc_map.end()) {
740
0
        return nullptr;
741
258k
    } else {
742
258k
        return i->second;
743
258k
    }
744
258k
}
745
746
0
std::string DescriptorTbl::debug_string() const {
747
0
    std::stringstream out;
748
0
    out << "tuples:\n";
749
750
0
    for (auto i : _tuple_desc_map) {
751
0
        out << i.second->debug_string() << '\n';
752
0
    }
753
754
0
    return out.str();
755
0
}
756
#include "common/compile_check_end.h"
757
} // namespace doris