/root/doris/be/src/runtime/descriptors.cpp
Line | Count | Source (jump to first uncovered line) |
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 | | |
29 | | #include <algorithm> |
30 | | #include <boost/algorithm/string/join.hpp> |
31 | | #include <memory> |
32 | | |
33 | | #include "common/object_pool.h" |
34 | | #include "runtime/primitive_type.h" |
35 | | #include "util/string_util.h" |
36 | | #include "vec/aggregate_functions/aggregate_function.h" |
37 | | #include "vec/data_types/data_type_factory.hpp" |
38 | | |
39 | | namespace doris { |
40 | | |
41 | | const int RowDescriptor::INVALID_IDX = -1; |
42 | | |
43 | | SlotDescriptor::SlotDescriptor(const TSlotDescriptor& tdesc) |
44 | | : _id(tdesc.id), |
45 | | _type(TypeDescriptor::from_thrift(tdesc.slotType)), |
46 | | _parent(tdesc.parent), |
47 | | _col_pos(tdesc.columnPos), |
48 | | _is_nullable(tdesc.nullIndicatorBit != -1), |
49 | | _col_name(tdesc.colName), |
50 | | _col_name_lower_case(to_lower(tdesc.colName)), |
51 | | _col_unique_id(tdesc.col_unique_id), |
52 | | _col_type(thrift_to_type(tdesc.primitive_type)), |
53 | | _slot_idx(tdesc.slotIdx), |
54 | | _field_idx(-1), |
55 | | _is_materialized(tdesc.isMaterialized), |
56 | | _is_key(tdesc.is_key), |
57 | | _need_materialize(tdesc.need_materialize), |
58 | | _column_paths(tdesc.column_paths), |
59 | | _is_auto_increment(tdesc.__isset.is_auto_increment ? tdesc.is_auto_increment : false), |
60 | 310 | _col_default_value(tdesc.__isset.col_default_value ? tdesc.col_default_value : "") {} |
61 | | |
62 | | SlotDescriptor::SlotDescriptor(const PSlotDescriptor& pdesc) |
63 | | : _id(pdesc.id()), |
64 | | _type(TypeDescriptor::from_protobuf(pdesc.slot_type())), |
65 | | _parent(pdesc.parent()), |
66 | | _col_pos(pdesc.column_pos()), |
67 | | _is_nullable(pdesc.null_indicator_bit() != -1), |
68 | | _col_name(pdesc.col_name()), |
69 | | _col_name_lower_case(to_lower(pdesc.col_name())), |
70 | | _col_unique_id(pdesc.col_unique_id()), |
71 | | _col_type(static_cast<PrimitiveType>(pdesc.col_type())), |
72 | | _slot_idx(pdesc.slot_idx()), |
73 | | _field_idx(-1), |
74 | | _is_materialized(pdesc.is_materialized()), |
75 | | _is_key(pdesc.is_key()), |
76 | | _need_materialize(true), |
77 | | _column_paths(pdesc.column_paths().begin(), pdesc.column_paths().end()), |
78 | 84 | _is_auto_increment(pdesc.is_auto_increment()) {} |
79 | | |
80 | 122 | void SlotDescriptor::to_protobuf(PSlotDescriptor* pslot) const { |
81 | 122 | pslot->set_id(_id); |
82 | 122 | pslot->set_parent(_parent); |
83 | 122 | _type.to_protobuf(pslot->mutable_slot_type()); |
84 | 122 | pslot->set_column_pos(_col_pos); |
85 | 122 | pslot->set_byte_offset(0); |
86 | 122 | pslot->set_null_indicator_byte(0); |
87 | 122 | pslot->set_null_indicator_bit(_is_nullable ? 0 : -1); |
88 | 122 | pslot->set_col_name(_col_name); |
89 | 122 | pslot->set_slot_idx(_slot_idx); |
90 | 122 | pslot->set_is_materialized(_is_materialized); |
91 | 122 | pslot->set_col_unique_id(_col_unique_id); |
92 | 122 | pslot->set_is_key(_is_key); |
93 | 122 | pslot->set_is_auto_increment(_is_auto_increment); |
94 | 122 | pslot->set_col_type(_col_type); |
95 | 122 | for (const std::string& path : _column_paths) { |
96 | 0 | pslot->add_column_paths(path); |
97 | 0 | } |
98 | 122 | } |
99 | | |
100 | 109 | vectorized::MutableColumnPtr SlotDescriptor::get_empty_mutable_column() const { |
101 | 109 | auto data_type = get_data_type_ptr(); |
102 | 109 | if (data_type) { |
103 | 109 | return data_type->create_column(); |
104 | 109 | } |
105 | 0 | return nullptr; |
106 | 109 | } |
107 | | |
108 | 254 | vectorized::DataTypePtr SlotDescriptor::get_data_type_ptr() const { |
109 | 254 | return vectorized::DataTypeFactory::instance().create_data_type(type(), is_nullable()); |
110 | 254 | } |
111 | | |
112 | 8 | std::string SlotDescriptor::debug_string() const { |
113 | 8 | std::stringstream out; |
114 | 8 | out << "Slot(id=" << _id << " type=" << _type << " col=" << _col_pos |
115 | 8 | << ", colname=" << _col_name << ", nullable=" << is_nullable() << ")"; |
116 | 8 | return out.str(); |
117 | 8 | } |
118 | | |
119 | | TableDescriptor::TableDescriptor(const TTableDescriptor& tdesc) |
120 | | : _table_type(tdesc.tableType), |
121 | | _name(tdesc.tableName), |
122 | | _database(tdesc.dbName), |
123 | | _table_id(tdesc.id), |
124 | | _num_cols(tdesc.numCols), |
125 | 3 | _num_clustering_cols(tdesc.numClusteringCols) {} |
126 | | |
127 | 0 | std::string TableDescriptor::debug_string() const { |
128 | 0 | std::stringstream out; |
129 | 0 | out << "#cols=" << _num_cols << " #clustering_cols=" << _num_clustering_cols; |
130 | 0 | return out.str(); |
131 | 0 | } |
132 | | |
133 | 3 | OlapTableDescriptor::OlapTableDescriptor(const TTableDescriptor& tdesc) : TableDescriptor(tdesc) {} |
134 | | |
135 | 0 | std::string OlapTableDescriptor::debug_string() const { |
136 | 0 | std::stringstream out; |
137 | 0 | out << "OlapTable(" << TableDescriptor::debug_string() << ")"; |
138 | 0 | return out.str(); |
139 | 0 | } |
140 | | |
141 | | SchemaTableDescriptor::SchemaTableDescriptor(const TTableDescriptor& tdesc) |
142 | 0 | : TableDescriptor(tdesc), _schema_table_type(tdesc.schemaTable.tableType) {} |
143 | 0 | SchemaTableDescriptor::~SchemaTableDescriptor() = default; |
144 | | |
145 | 0 | std::string SchemaTableDescriptor::debug_string() const { |
146 | 0 | std::stringstream out; |
147 | 0 | out << "SchemaTable(" << TableDescriptor::debug_string() << ")"; |
148 | 0 | return out.str(); |
149 | 0 | } |
150 | | |
151 | | BrokerTableDescriptor::BrokerTableDescriptor(const TTableDescriptor& tdesc) |
152 | 0 | : TableDescriptor(tdesc) {} |
153 | | |
154 | 0 | BrokerTableDescriptor::~BrokerTableDescriptor() = default; |
155 | | |
156 | 0 | std::string BrokerTableDescriptor::debug_string() const { |
157 | 0 | std::stringstream out; |
158 | 0 | out << "BrokerTable(" << TableDescriptor::debug_string() << ")"; |
159 | 0 | return out.str(); |
160 | 0 | } |
161 | | |
162 | 0 | HiveTableDescriptor::HiveTableDescriptor(const TTableDescriptor& tdesc) : TableDescriptor(tdesc) {} |
163 | | |
164 | 0 | HiveTableDescriptor::~HiveTableDescriptor() = default; |
165 | | |
166 | 0 | std::string HiveTableDescriptor::debug_string() const { |
167 | 0 | std::stringstream out; |
168 | 0 | out << "HiveTable(" << TableDescriptor::debug_string() << ")"; |
169 | 0 | return out.str(); |
170 | 0 | } |
171 | | |
172 | | IcebergTableDescriptor::IcebergTableDescriptor(const TTableDescriptor& tdesc) |
173 | 0 | : TableDescriptor(tdesc) {} |
174 | | |
175 | 0 | IcebergTableDescriptor::~IcebergTableDescriptor() = default; |
176 | | |
177 | 0 | std::string IcebergTableDescriptor::debug_string() const { |
178 | 0 | std::stringstream out; |
179 | 0 | out << "IcebergTable(" << TableDescriptor::debug_string() << ")"; |
180 | 0 | return out.str(); |
181 | 0 | } |
182 | | |
183 | | MaxComputeTableDescriptor::MaxComputeTableDescriptor(const TTableDescriptor& tdesc) |
184 | | : TableDescriptor(tdesc), |
185 | | _region(tdesc.mcTable.region), |
186 | | _project(tdesc.mcTable.project), |
187 | | _table(tdesc.mcTable.table), |
188 | | _odps_url(tdesc.mcTable.odps_url), |
189 | | _tunnel_url(tdesc.mcTable.tunnel_url), |
190 | | _access_key(tdesc.mcTable.access_key), |
191 | | _secret_key(tdesc.mcTable.secret_key), |
192 | 0 | _public_access(tdesc.mcTable.public_access) { |
193 | 0 | if (tdesc.mcTable.__isset.endpoint) { |
194 | 0 | _endpoint = tdesc.mcTable.endpoint; |
195 | 0 | } else { |
196 | 0 | _init_status = Status::InvalidArgument( |
197 | 0 | "fail to init MaxComputeTableDescriptor, missing endpoint."); |
198 | 0 | } |
199 | |
|
200 | 0 | if (tdesc.mcTable.__isset.quota) { |
201 | 0 | _quota = tdesc.mcTable.quota; |
202 | 0 | } else { |
203 | 0 | _init_status = |
204 | 0 | Status::InvalidArgument("fail to init MaxComputeTableDescriptor, missing quota."); |
205 | 0 | } |
206 | 0 | } |
207 | | |
208 | 0 | MaxComputeTableDescriptor::~MaxComputeTableDescriptor() = default; |
209 | | |
210 | 0 | std::string MaxComputeTableDescriptor::debug_string() const { |
211 | 0 | std::stringstream out; |
212 | 0 | out << "MaxComputeTable(" << TableDescriptor::debug_string() << ")"; |
213 | 0 | return out.str(); |
214 | 0 | } |
215 | | |
216 | 0 | EsTableDescriptor::EsTableDescriptor(const TTableDescriptor& tdesc) : TableDescriptor(tdesc) {} |
217 | | |
218 | 0 | EsTableDescriptor::~EsTableDescriptor() = default; |
219 | | |
220 | 0 | std::string EsTableDescriptor::debug_string() const { |
221 | 0 | std::stringstream out; |
222 | 0 | out << "EsTable(" << TableDescriptor::debug_string() << ")"; |
223 | 0 | return out.str(); |
224 | 0 | } |
225 | | |
226 | | MySQLTableDescriptor::MySQLTableDescriptor(const TTableDescriptor& tdesc) |
227 | | : TableDescriptor(tdesc), |
228 | | _mysql_db(tdesc.mysqlTable.db), |
229 | | _mysql_table(tdesc.mysqlTable.table), |
230 | | _host(tdesc.mysqlTable.host), |
231 | | _port(tdesc.mysqlTable.port), |
232 | | _user(tdesc.mysqlTable.user), |
233 | | _passwd(tdesc.mysqlTable.passwd), |
234 | 0 | _charset(tdesc.mysqlTable.charset) {} |
235 | | |
236 | 0 | std::string MySQLTableDescriptor::debug_string() const { |
237 | 0 | std::stringstream out; |
238 | 0 | out << "MySQLTable(" << TableDescriptor::debug_string() << " _db" << _mysql_db |
239 | 0 | << " table=" << _mysql_table << " host=" << _host << " port=" << _port << " user=" << _user |
240 | 0 | << " passwd=" << _passwd << " charset=" << _charset; |
241 | 0 | return out.str(); |
242 | 0 | } |
243 | | |
244 | | ODBCTableDescriptor::ODBCTableDescriptor(const TTableDescriptor& tdesc) |
245 | | : TableDescriptor(tdesc), |
246 | | _db(tdesc.odbcTable.db), |
247 | | _table(tdesc.odbcTable.table), |
248 | | _host(tdesc.odbcTable.host), |
249 | | _port(tdesc.odbcTable.port), |
250 | | _user(tdesc.odbcTable.user), |
251 | | _passwd(tdesc.odbcTable.passwd), |
252 | | _driver(tdesc.odbcTable.driver), |
253 | 0 | _type(tdesc.odbcTable.type) {} |
254 | | |
255 | 0 | std::string ODBCTableDescriptor::debug_string() const { |
256 | 0 | std::stringstream out; |
257 | 0 | out << "ODBCTable(" << TableDescriptor::debug_string() << " _db" << _db << " table=" << _table |
258 | 0 | << " host=" << _host << " port=" << _port << " user=" << _user << " passwd=" << _passwd |
259 | 0 | << " driver=" << _driver << " type" << _type; |
260 | 0 | return out.str(); |
261 | 0 | } |
262 | | |
263 | | JdbcTableDescriptor::JdbcTableDescriptor(const TTableDescriptor& tdesc) |
264 | | : TableDescriptor(tdesc), |
265 | | _jdbc_catalog_id(tdesc.jdbcTable.catalog_id), |
266 | | _jdbc_resource_name(tdesc.jdbcTable.jdbc_resource_name), |
267 | | _jdbc_driver_url(tdesc.jdbcTable.jdbc_driver_url), |
268 | | _jdbc_driver_class(tdesc.jdbcTable.jdbc_driver_class), |
269 | | _jdbc_driver_checksum(tdesc.jdbcTable.jdbc_driver_checksum), |
270 | | _jdbc_url(tdesc.jdbcTable.jdbc_url), |
271 | | _jdbc_table_name(tdesc.jdbcTable.jdbc_table_name), |
272 | | _jdbc_user(tdesc.jdbcTable.jdbc_user), |
273 | | _jdbc_passwd(tdesc.jdbcTable.jdbc_password), |
274 | | _connection_pool_min_size(tdesc.jdbcTable.connection_pool_min_size), |
275 | | _connection_pool_max_size(tdesc.jdbcTable.connection_pool_max_size), |
276 | | _connection_pool_max_wait_time(tdesc.jdbcTable.connection_pool_max_wait_time), |
277 | | _connection_pool_max_life_time(tdesc.jdbcTable.connection_pool_max_life_time), |
278 | 0 | _connection_pool_keep_alive(tdesc.jdbcTable.connection_pool_keep_alive) {} |
279 | | |
280 | 0 | std::string JdbcTableDescriptor::debug_string() const { |
281 | 0 | fmt::memory_buffer buf; |
282 | 0 | fmt::format_to( |
283 | 0 | buf, |
284 | 0 | "JDBCTable({} ,_jdbc_catalog_id = {}, _jdbc_resource_name={} ,_jdbc_driver_url={} " |
285 | 0 | ",_jdbc_driver_class={} ,_jdbc_driver_checksum={} ,_jdbc_url={} " |
286 | 0 | ",_jdbc_table_name={} ,_jdbc_user={} ,_jdbc_passwd={} ,_connection_pool_min_size={} " |
287 | 0 | ",_connection_pool_max_size={} ,_connection_pool_max_wait_time={} " |
288 | 0 | ",_connection_pool_max_life_time={} ,_connection_pool_keep_alive={})", |
289 | 0 | TableDescriptor::debug_string(), _jdbc_catalog_id, _jdbc_resource_name, |
290 | 0 | _jdbc_driver_url, _jdbc_driver_class, _jdbc_driver_checksum, _jdbc_url, |
291 | 0 | _jdbc_table_name, _jdbc_user, _jdbc_passwd, _connection_pool_min_size, |
292 | 0 | _connection_pool_max_size, _connection_pool_max_wait_time, |
293 | 0 | _connection_pool_max_life_time, _connection_pool_keep_alive); |
294 | 0 | return fmt::to_string(buf); |
295 | 0 | } |
296 | | |
297 | | TupleDescriptor::TupleDescriptor(const TTupleDescriptor& tdesc, bool own_slots) |
298 | | : _id(tdesc.id), |
299 | | _num_materialized_slots(0), |
300 | | _has_varlen_slots(false), |
301 | 55 | _own_slots(own_slots) {} |
302 | | |
303 | | TupleDescriptor::TupleDescriptor(const PTupleDescriptor& pdesc, bool own_slots) |
304 | | : _id(pdesc.id()), |
305 | | _num_materialized_slots(0), |
306 | | _has_varlen_slots(false), |
307 | 21 | _own_slots(own_slots) {} |
308 | | |
309 | 394 | void TupleDescriptor::add_slot(SlotDescriptor* slot) { |
310 | 394 | _slots.push_back(slot); |
311 | | |
312 | 394 | if (slot->is_materialized()) { |
313 | 366 | ++_num_materialized_slots; |
314 | | |
315 | 366 | if (slot->type().is_string_type() || slot->type().is_collection_type()) { |
316 | 99 | _has_varlen_slots = true; |
317 | 99 | } |
318 | 366 | } |
319 | 394 | } |
320 | | |
321 | 21 | void TupleDescriptor::to_protobuf(PTupleDescriptor* ptuple) const { |
322 | 21 | ptuple->Clear(); |
323 | 21 | ptuple->set_id(_id); |
324 | | // Useless not set |
325 | 21 | ptuple->set_byte_size(0); |
326 | 21 | ptuple->set_table_id(-1); |
327 | 21 | ptuple->set_num_null_bytes(0); |
328 | 21 | } |
329 | | |
330 | 3 | std::string TupleDescriptor::debug_string() const { |
331 | 3 | std::stringstream out; |
332 | 3 | out << "Tuple(id=" << _id; |
333 | 3 | if (_table_desc != nullptr) { |
334 | | //out << " " << _table_desc->debug_string(); |
335 | 0 | } |
336 | | |
337 | 3 | out << " slots=["; |
338 | 11 | for (size_t i = 0; i < _slots.size(); ++i) { |
339 | 8 | if (i > 0) { |
340 | 5 | out << ", "; |
341 | 5 | } |
342 | 8 | out << _slots[i]->debug_string(); |
343 | 8 | } |
344 | | |
345 | 3 | out << "]"; |
346 | 3 | out << " has_varlen_slots=" << _has_varlen_slots; |
347 | 3 | out << ")"; |
348 | 3 | return out.str(); |
349 | 3 | } |
350 | | |
351 | | RowDescriptor::RowDescriptor(const DescriptorTbl& desc_tbl, const std::vector<TTupleId>& row_tuples, |
352 | | const std::vector<bool>& nullable_tuples) |
353 | 10 | : _tuple_idx_nullable_map(nullable_tuples) { |
354 | 10 | DCHECK(nullable_tuples.size() == row_tuples.size()) |
355 | 0 | << "nullable_tuples size " << nullable_tuples.size() << " != row_tuples size " |
356 | 0 | << row_tuples.size(); |
357 | 10 | DCHECK_GT(row_tuples.size(), 0); |
358 | 10 | _num_materialized_slots = 0; |
359 | 10 | _num_slots = 0; |
360 | | |
361 | 10 | for (int row_tuple : row_tuples) { |
362 | 10 | TupleDescriptor* tupleDesc = desc_tbl.get_tuple_descriptor(row_tuple); |
363 | 10 | _num_materialized_slots += tupleDesc->num_materialized_slots(); |
364 | 10 | _num_slots += tupleDesc->slots().size(); |
365 | 10 | _tuple_desc_map.push_back(tupleDesc); |
366 | 10 | DCHECK(_tuple_desc_map.back() != nullptr); |
367 | 10 | } |
368 | | |
369 | 10 | init_tuple_idx_map(); |
370 | 10 | init_has_varlen_slots(); |
371 | 10 | } |
372 | | |
373 | | RowDescriptor::RowDescriptor(TupleDescriptor* tuple_desc, bool is_nullable) |
374 | 33 | : _tuple_desc_map(1, tuple_desc), _tuple_idx_nullable_map(1, is_nullable) { |
375 | 33 | init_tuple_idx_map(); |
376 | 33 | init_has_varlen_slots(); |
377 | 33 | _num_slots = tuple_desc->slots().size(); |
378 | 33 | } |
379 | | |
380 | 0 | RowDescriptor::RowDescriptor(const RowDescriptor& lhs_row_desc, const RowDescriptor& rhs_row_desc) { |
381 | 0 | _tuple_desc_map.insert(_tuple_desc_map.end(), lhs_row_desc._tuple_desc_map.begin(), |
382 | 0 | lhs_row_desc._tuple_desc_map.end()); |
383 | 0 | _tuple_desc_map.insert(_tuple_desc_map.end(), rhs_row_desc._tuple_desc_map.begin(), |
384 | 0 | rhs_row_desc._tuple_desc_map.end()); |
385 | 0 | _tuple_idx_nullable_map.insert(_tuple_idx_nullable_map.end(), |
386 | 0 | lhs_row_desc._tuple_idx_nullable_map.begin(), |
387 | 0 | lhs_row_desc._tuple_idx_nullable_map.end()); |
388 | 0 | _tuple_idx_nullable_map.insert(_tuple_idx_nullable_map.end(), |
389 | 0 | rhs_row_desc._tuple_idx_nullable_map.begin(), |
390 | 0 | rhs_row_desc._tuple_idx_nullable_map.end()); |
391 | 0 | init_tuple_idx_map(); |
392 | 0 | init_has_varlen_slots(); |
393 | |
|
394 | 0 | _num_slots = lhs_row_desc.num_slots() + rhs_row_desc.num_slots(); |
395 | 0 | } |
396 | | |
397 | 43 | void RowDescriptor::init_tuple_idx_map() { |
398 | | // find max id |
399 | 43 | TupleId max_id = 0; |
400 | 43 | for (auto& i : _tuple_desc_map) { |
401 | 43 | max_id = std::max(i->id(), max_id); |
402 | 43 | } |
403 | | |
404 | 43 | _tuple_idx_map.resize(max_id + 1, INVALID_IDX); |
405 | 86 | for (int i = 0; i < _tuple_desc_map.size(); ++i) { |
406 | 43 | _tuple_idx_map[_tuple_desc_map[i]->id()] = i; |
407 | 43 | } |
408 | 43 | } |
409 | | |
410 | 43 | void RowDescriptor::init_has_varlen_slots() { |
411 | 43 | _has_varlen_slots = false; |
412 | 43 | for (auto& i : _tuple_desc_map) { |
413 | 43 | if (i->has_varlen_slots()) { |
414 | 26 | _has_varlen_slots = true; |
415 | 26 | break; |
416 | 26 | } |
417 | 43 | } |
418 | 43 | } |
419 | | |
420 | 0 | int RowDescriptor::get_tuple_idx(TupleId id) const { |
421 | | // comment CHECK temporarily to make fuzzy test run smoothly |
422 | | // DCHECK_LT(id, _tuple_idx_map.size()) << "RowDescriptor: " << debug_string(); |
423 | 0 | if (_tuple_idx_map.size() <= id) { |
424 | 0 | return RowDescriptor::INVALID_IDX; |
425 | 0 | } |
426 | 0 | return _tuple_idx_map[id]; |
427 | 0 | } |
428 | | |
429 | 0 | void RowDescriptor::to_thrift(std::vector<TTupleId>* row_tuple_ids) { |
430 | 0 | row_tuple_ids->clear(); |
431 | |
|
432 | 0 | for (auto& i : _tuple_desc_map) { |
433 | 0 | row_tuple_ids->push_back(i->id()); |
434 | 0 | } |
435 | 0 | } |
436 | | |
437 | | void RowDescriptor::to_protobuf( |
438 | 0 | google::protobuf::RepeatedField<google::protobuf::int32>* row_tuple_ids) const { |
439 | 0 | row_tuple_ids->Clear(); |
440 | 0 | for (auto* desc : _tuple_desc_map) { |
441 | 0 | row_tuple_ids->Add(desc->id()); |
442 | 0 | } |
443 | 0 | } |
444 | | |
445 | 0 | bool RowDescriptor::is_prefix_of(const RowDescriptor& other_desc) const { |
446 | 0 | if (_tuple_desc_map.size() > other_desc._tuple_desc_map.size()) { |
447 | 0 | return false; |
448 | 0 | } |
449 | | |
450 | 0 | for (int i = 0; i < _tuple_desc_map.size(); ++i) { |
451 | | // pointer comparison okay, descriptors are unique |
452 | 0 | if (_tuple_desc_map[i] != other_desc._tuple_desc_map[i]) { |
453 | 0 | return false; |
454 | 0 | } |
455 | 0 | } |
456 | | |
457 | 0 | return true; |
458 | 0 | } |
459 | | |
460 | 0 | bool RowDescriptor::equals(const RowDescriptor& other_desc) const { |
461 | 0 | if (_tuple_desc_map.size() != other_desc._tuple_desc_map.size()) { |
462 | 0 | return false; |
463 | 0 | } |
464 | | |
465 | 0 | for (int i = 0; i < _tuple_desc_map.size(); ++i) { |
466 | | // pointer comparison okay, descriptors are unique |
467 | 0 | if (_tuple_desc_map[i] != other_desc._tuple_desc_map[i]) { |
468 | 0 | return false; |
469 | 0 | } |
470 | 0 | } |
471 | | |
472 | 0 | return true; |
473 | 0 | } |
474 | | |
475 | 0 | std::string RowDescriptor::debug_string() const { |
476 | 0 | std::stringstream ss; |
477 | |
|
478 | 0 | ss << "tuple_desc_map: ["; |
479 | 0 | for (int i = 0; i < _tuple_desc_map.size(); ++i) { |
480 | 0 | ss << _tuple_desc_map[i]->debug_string(); |
481 | 0 | if (i != _tuple_desc_map.size() - 1) { |
482 | 0 | ss << ", "; |
483 | 0 | } |
484 | 0 | } |
485 | 0 | ss << "] "; |
486 | |
|
487 | 0 | ss << "tuple_id_map: ["; |
488 | 0 | for (int i = 0; i < _tuple_idx_map.size(); ++i) { |
489 | 0 | ss << _tuple_idx_map[i]; |
490 | 0 | if (i != _tuple_idx_map.size() - 1) { |
491 | 0 | ss << ", "; |
492 | 0 | } |
493 | 0 | } |
494 | 0 | ss << "] "; |
495 | |
|
496 | 0 | ss << "tuple_is_nullable: ["; |
497 | 0 | for (int i = 0; i < _tuple_idx_nullable_map.size(); ++i) { |
498 | 0 | ss << _tuple_idx_nullable_map[i]; |
499 | 0 | if (i != _tuple_idx_nullable_map.size() - 1) { |
500 | 0 | ss << ", "; |
501 | 0 | } |
502 | 0 | } |
503 | 0 | ss << "] "; |
504 | |
|
505 | 0 | return ss.str(); |
506 | 0 | } |
507 | | |
508 | 16 | int RowDescriptor::get_column_id(int slot_id, bool force_materialize_slot) const { |
509 | 16 | int column_id_counter = 0; |
510 | 16 | for (auto* const tuple_desc : _tuple_desc_map) { |
511 | 22 | for (auto* const slot : tuple_desc->slots()) { |
512 | 22 | if (!force_materialize_slot && !slot->need_materialize()) { |
513 | 0 | continue; |
514 | 0 | } |
515 | 22 | if (slot->id() == slot_id) { |
516 | 16 | return column_id_counter; |
517 | 16 | } |
518 | 6 | column_id_counter++; |
519 | 6 | } |
520 | 16 | } |
521 | 0 | return -1; |
522 | 16 | } |
523 | | |
524 | | Status DescriptorTbl::create(ObjectPool* pool, const TDescriptorTable& thrift_tbl, |
525 | 27 | DescriptorTbl** tbl) { |
526 | 27 | *tbl = pool->add(new DescriptorTbl()); |
527 | | |
528 | | // deserialize table descriptors first, they are being referenced by tuple descriptors |
529 | 27 | for (const auto& tdesc : thrift_tbl.tableDescriptors) { |
530 | 3 | TableDescriptor* desc = nullptr; |
531 | | |
532 | 3 | switch (tdesc.tableType) { |
533 | 0 | case TTableType::MYSQL_TABLE: |
534 | 0 | desc = pool->add(new MySQLTableDescriptor(tdesc)); |
535 | 0 | break; |
536 | | |
537 | 0 | case TTableType::ODBC_TABLE: |
538 | 0 | desc = pool->add(new ODBCTableDescriptor(tdesc)); |
539 | 0 | break; |
540 | | |
541 | 3 | case TTableType::OLAP_TABLE: |
542 | 3 | desc = pool->add(new OlapTableDescriptor(tdesc)); |
543 | 3 | break; |
544 | | |
545 | 0 | case TTableType::SCHEMA_TABLE: |
546 | 0 | desc = pool->add(new SchemaTableDescriptor(tdesc)); |
547 | 0 | break; |
548 | 0 | case TTableType::BROKER_TABLE: |
549 | 0 | desc = pool->add(new BrokerTableDescriptor(tdesc)); |
550 | 0 | break; |
551 | 0 | case TTableType::ES_TABLE: |
552 | 0 | desc = pool->add(new EsTableDescriptor(tdesc)); |
553 | 0 | break; |
554 | 0 | case TTableType::HIVE_TABLE: |
555 | 0 | desc = pool->add(new HiveTableDescriptor(tdesc)); |
556 | 0 | break; |
557 | 0 | case TTableType::ICEBERG_TABLE: |
558 | 0 | desc = pool->add(new IcebergTableDescriptor(tdesc)); |
559 | 0 | break; |
560 | 0 | case TTableType::JDBC_TABLE: |
561 | 0 | desc = pool->add(new JdbcTableDescriptor(tdesc)); |
562 | 0 | break; |
563 | 0 | case TTableType::MAX_COMPUTE_TABLE: |
564 | 0 | desc = pool->add(new MaxComputeTableDescriptor(tdesc)); |
565 | 0 | break; |
566 | 0 | default: |
567 | 0 | DCHECK(false) << "invalid table type: " << tdesc.tableType; |
568 | 3 | } |
569 | | |
570 | 3 | (*tbl)->_tbl_desc_map[tdesc.id] = desc; |
571 | 3 | } |
572 | | |
573 | 31 | for (const auto& tdesc : thrift_tbl.tupleDescriptors) { |
574 | 31 | TupleDescriptor* desc = pool->add(new TupleDescriptor(tdesc)); |
575 | | |
576 | | // fix up table pointer |
577 | 31 | if (tdesc.__isset.tableId) { |
578 | 3 | desc->_table_desc = (*tbl)->get_table_descriptor(tdesc.tableId); |
579 | 3 | DCHECK(desc->_table_desc != nullptr); |
580 | 3 | } |
581 | | |
582 | 31 | (*tbl)->_tuple_desc_map[tdesc.id] = desc; |
583 | 31 | (*tbl)->_row_tuples.emplace_back(tdesc.id); |
584 | 31 | } |
585 | | |
586 | 127 | for (const auto& tdesc : thrift_tbl.slotDescriptors) { |
587 | 127 | SlotDescriptor* slot_d = pool->add(new SlotDescriptor(tdesc)); |
588 | 127 | (*tbl)->_slot_desc_map[tdesc.id] = slot_d; |
589 | | |
590 | | // link to parent |
591 | 127 | auto entry = (*tbl)->_tuple_desc_map.find(tdesc.parent); |
592 | | |
593 | 127 | if (entry == (*tbl)->_tuple_desc_map.end()) { |
594 | 0 | return Status::InternalError("unknown tid in slot descriptor msg"); |
595 | 0 | } |
596 | 127 | entry->second->add_slot(slot_d); |
597 | 127 | } |
598 | | |
599 | 27 | return Status::OK(); |
600 | 27 | } |
601 | | |
602 | 3 | TableDescriptor* DescriptorTbl::get_table_descriptor(TableId id) const { |
603 | | // TODO: is there some boost function to do exactly this? |
604 | 3 | auto i = _tbl_desc_map.find(id); |
605 | | |
606 | 3 | if (i == _tbl_desc_map.end()) { |
607 | 0 | return nullptr; |
608 | 3 | } else { |
609 | 3 | return i->second; |
610 | 3 | } |
611 | 3 | } |
612 | | |
613 | 46 | TupleDescriptor* DescriptorTbl::get_tuple_descriptor(TupleId id) const { |
614 | | // TODO: is there some boost function to do exactly this? |
615 | 46 | auto i = _tuple_desc_map.find(id); |
616 | | |
617 | 46 | if (i == _tuple_desc_map.end()) { |
618 | 0 | return nullptr; |
619 | 46 | } else { |
620 | 46 | return i->second; |
621 | 46 | } |
622 | 46 | } |
623 | | |
624 | 16 | SlotDescriptor* DescriptorTbl::get_slot_descriptor(SlotId id) const { |
625 | | // TODO: is there some boost function to do exactly this? |
626 | 16 | auto i = _slot_desc_map.find(id); |
627 | | |
628 | 16 | if (i == _slot_desc_map.end()) { |
629 | 0 | return nullptr; |
630 | 16 | } else { |
631 | 16 | return i->second; |
632 | 16 | } |
633 | 16 | } |
634 | | |
635 | 0 | std::string DescriptorTbl::debug_string() const { |
636 | 0 | std::stringstream out; |
637 | 0 | out << "tuples:\n"; |
638 | |
|
639 | 0 | for (auto i : _tuple_desc_map) { |
640 | 0 | out << i.second->debug_string() << '\n'; |
641 | 0 | } |
642 | |
|
643 | 0 | return out.str(); |
644 | 0 | } |
645 | | |
646 | | } // namespace doris |