-
Notifications
You must be signed in to change notification settings - Fork 114
/
Copy pathstatement.cc
1429 lines (1277 loc) · 51.9 KB
/
statement.cc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements. See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership. The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License. You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing,
// software distributed under the License is distributed on an
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied. See the License for the
// specific language governing permissions and limitations
// under the License.
#include "statement.h"
#include <array>
#include <cassert>
#include <cerrno>
#include <cinttypes>
#include <cstring>
#include <iostream>
#include <memory>
#include <utility>
#include <vector>
#include <adbc.h>
#include <libpq-fe.h>
#include <nanoarrow/nanoarrow.hpp>
#include "common/options.h"
#include "common/utils.h"
#include "connection.h"
#include "error.h"
#include "postgres_copy_reader.h"
#include "postgres_type.h"
#include "postgres_util.h"
#include "result_helper.h"
namespace adbcpq {
namespace {
/// The flag indicating to PostgreSQL that we want binary-format values.
constexpr int kPgBinaryFormat = 1;
/// One-value ArrowArrayStream used to unify the implementations of Bind
struct OneValueStream {
struct ArrowSchema schema;
struct ArrowArray array;
static int GetSchema(struct ArrowArrayStream* self, struct ArrowSchema* out) {
OneValueStream* stream = static_cast<OneValueStream*>(self->private_data);
return ArrowSchemaDeepCopy(&stream->schema, out);
}
static int GetNext(struct ArrowArrayStream* self, struct ArrowArray* out) {
OneValueStream* stream = static_cast<OneValueStream*>(self->private_data);
*out = stream->array;
stream->array.release = nullptr;
return 0;
}
static const char* GetLastError(struct ArrowArrayStream* self) { return NULL; }
static void Release(struct ArrowArrayStream* self) {
OneValueStream* stream = static_cast<OneValueStream*>(self->private_data);
if (stream->schema.release) {
stream->schema.release(&stream->schema);
stream->schema.release = nullptr;
}
if (stream->array.release) {
stream->array.release(&stream->array);
stream->array.release = nullptr;
}
delete stream;
self->release = nullptr;
}
};
/// Helper to manage resources with RAII
template <typename T>
struct Releaser {
static void Release(T* value) {
if (value->release) {
value->release(value);
}
}
};
template <>
struct Releaser<struct ArrowArrayView> {
static void Release(struct ArrowArrayView* value) {
if (value->storage_type != NANOARROW_TYPE_UNINITIALIZED) {
ArrowArrayViewReset(value);
}
}
};
template <typename Resource>
struct Handle {
Resource value;
Handle() { std::memset(&value, 0, sizeof(value)); }
~Handle() { Releaser<Resource>::Release(&value); }
Resource* operator->() { return &value; }
};
/// Build an PostgresType object from a PGresult*
AdbcStatusCode ResolvePostgresType(const PostgresTypeResolver& type_resolver,
PGresult* result, PostgresType* out,
struct AdbcError* error) {
ArrowError na_error;
const int num_fields = PQnfields(result);
PostgresType root_type(PostgresTypeId::kRecord);
for (int i = 0; i < num_fields; i++) {
const Oid pg_oid = PQftype(result, i);
PostgresType pg_type;
if (type_resolver.Find(pg_oid, &pg_type, &na_error) != NANOARROW_OK) {
SetError(error, "%s%d%s%s%s%d", "[libpq] Column #", i + 1, " (\"",
PQfname(result, i), "\") has unknown type code ", pg_oid);
return ADBC_STATUS_NOT_IMPLEMENTED;
}
root_type.AppendChild(PQfname(result, i), pg_type);
}
*out = root_type;
return ADBC_STATUS_OK;
}
/// Helper to manage bind parameters with a prepared statement
struct BindStream {
Handle<struct ArrowArrayStream> bind;
Handle<struct ArrowSchema> bind_schema;
struct ArrowSchemaView bind_schema_view;
std::vector<struct ArrowSchemaView> bind_schema_fields;
// OIDs for parameter types
std::vector<uint32_t> param_types;
std::vector<char*> param_values;
std::vector<int> param_lengths;
std::vector<int> param_formats;
std::vector<size_t> param_values_offsets;
std::vector<char> param_values_buffer;
// XXX: this assumes fixed-length fields only - will need more
// consideration to deal with variable-length fields
bool has_tz_field = false;
std::string tz_setting;
struct ArrowError na_error;
explicit BindStream(struct ArrowArrayStream&& bind) {
this->bind.value = std::move(bind);
std::memset(&na_error, 0, sizeof(na_error));
}
template <typename Callback>
AdbcStatusCode Begin(Callback&& callback, struct AdbcError* error) {
CHECK_NA(INTERNAL, bind->get_schema(&bind.value, &bind_schema.value), error);
CHECK_NA(
INTERNAL,
ArrowSchemaViewInit(&bind_schema_view, &bind_schema.value, /*error*/ nullptr),
error);
if (bind_schema_view.type != ArrowType::NANOARROW_TYPE_STRUCT) {
SetError(error, "%s", "[libpq] Bind parameters must have type STRUCT");
return ADBC_STATUS_INVALID_STATE;
}
bind_schema_fields.resize(bind_schema->n_children);
for (size_t i = 0; i < bind_schema_fields.size(); i++) {
CHECK_NA(INTERNAL,
ArrowSchemaViewInit(&bind_schema_fields[i], bind_schema->children[i],
/*error*/ nullptr),
error);
}
return std::move(callback)();
}
AdbcStatusCode SetParamTypes(const PostgresTypeResolver& type_resolver,
struct AdbcError* error) {
param_types.resize(bind_schema->n_children);
param_values.resize(bind_schema->n_children);
param_lengths.resize(bind_schema->n_children);
param_formats.resize(bind_schema->n_children, kPgBinaryFormat);
param_values_offsets.reserve(bind_schema->n_children);
for (size_t i = 0; i < bind_schema_fields.size(); i++) {
PostgresTypeId type_id;
switch (bind_schema_fields[i].type) {
case ArrowType::NANOARROW_TYPE_BOOL:
type_id = PostgresTypeId::kBool;
param_lengths[i] = 1;
break;
case ArrowType::NANOARROW_TYPE_INT8:
case ArrowType::NANOARROW_TYPE_INT16:
type_id = PostgresTypeId::kInt2;
param_lengths[i] = 2;
break;
case ArrowType::NANOARROW_TYPE_INT32:
type_id = PostgresTypeId::kInt4;
param_lengths[i] = 4;
break;
case ArrowType::NANOARROW_TYPE_INT64:
type_id = PostgresTypeId::kInt8;
param_lengths[i] = 8;
break;
case ArrowType::NANOARROW_TYPE_FLOAT:
type_id = PostgresTypeId::kFloat4;
param_lengths[i] = 4;
break;
case ArrowType::NANOARROW_TYPE_DOUBLE:
type_id = PostgresTypeId::kFloat8;
param_lengths[i] = 8;
break;
case ArrowType::NANOARROW_TYPE_STRING:
case ArrowType::NANOARROW_TYPE_LARGE_STRING:
type_id = PostgresTypeId::kText;
param_lengths[i] = 0;
break;
case ArrowType::NANOARROW_TYPE_BINARY:
type_id = PostgresTypeId::kBytea;
param_lengths[i] = 0;
break;
case ArrowType::NANOARROW_TYPE_DATE32:
type_id = PostgresTypeId::kDate;
param_lengths[i] = 4;
break;
case ArrowType::NANOARROW_TYPE_TIMESTAMP:
type_id = PostgresTypeId::kTimestamp;
param_lengths[i] = 8;
break;
case ArrowType::NANOARROW_TYPE_DURATION:
case ArrowType::NANOARROW_TYPE_INTERVAL_MONTH_DAY_NANO:
type_id = PostgresTypeId::kInterval;
param_lengths[i] = 16;
break;
default:
SetError(error, "%s%" PRIu64 "%s%s%s%s", "[libpq] Field #",
static_cast<uint64_t>(i + 1), " ('", bind_schema->children[i]->name,
"') has unsupported parameter type ",
ArrowTypeString(bind_schema_fields[i].type));
return ADBC_STATUS_NOT_IMPLEMENTED;
}
param_types[i] = type_resolver.GetOID(type_id);
if (param_types[i] == 0) {
SetError(error, "%s%" PRIu64 "%s%s%s%s", "[libpq] Field #",
static_cast<uint64_t>(i + 1), " ('", bind_schema->children[i]->name,
"') has type with no corresponding PostgreSQL type ",
ArrowTypeString(bind_schema_fields[i].type));
return ADBC_STATUS_NOT_IMPLEMENTED;
}
}
size_t param_values_length = 0;
for (int length : param_lengths) {
param_values_offsets.push_back(param_values_length);
param_values_length += length;
}
param_values_buffer.resize(param_values_length);
return ADBC_STATUS_OK;
}
AdbcStatusCode Prepare(PGconn* conn, const std::string& query, struct AdbcError* error,
const bool autocommit) {
// tz-aware timestamps require special handling to set the timezone to UTC
// prior to sending over the binary protocol; must be reset after execute
for (int64_t col = 0; col < bind_schema->n_children; col++) {
if ((bind_schema_fields[col].type == ArrowType::NANOARROW_TYPE_TIMESTAMP) &&
(strcmp("", bind_schema_fields[col].timezone))) {
has_tz_field = true;
if (autocommit) {
PGresult* begin_result = PQexec(conn, "BEGIN");
if (PQresultStatus(begin_result) != PGRES_COMMAND_OK) {
AdbcStatusCode code =
SetError(error, begin_result,
"[libpq] Failed to begin transaction for timezone data: %s",
PQerrorMessage(conn));
PQclear(begin_result);
return code;
}
PQclear(begin_result);
}
PGresult* get_tz_result = PQexec(conn, "SELECT current_setting('TIMEZONE')");
if (PQresultStatus(get_tz_result) != PGRES_TUPLES_OK) {
AdbcStatusCode code = SetError(error, get_tz_result,
"[libpq] Could not query current timezone: %s",
PQerrorMessage(conn));
PQclear(get_tz_result);
return code;
}
tz_setting = std::string(PQgetvalue(get_tz_result, 0, 0));
PQclear(get_tz_result);
PGresult* set_utc_result = PQexec(conn, "SET TIME ZONE 'UTC'");
if (PQresultStatus(set_utc_result) != PGRES_COMMAND_OK) {
AdbcStatusCode code = SetError(error, set_utc_result,
"[libpq] Failed to set time zone to UTC: %s",
PQerrorMessage(conn));
PQclear(set_utc_result);
return code;
}
PQclear(set_utc_result);
break;
}
}
PGresult* result = PQprepare(conn, /*stmtName=*/"", query.c_str(),
/*nParams=*/bind_schema->n_children, param_types.data());
if (PQresultStatus(result) != PGRES_COMMAND_OK) {
AdbcStatusCode code =
SetError(error, result, "[libpq] Failed to prepare query: %s\nQuery was:%s",
PQerrorMessage(conn), query.c_str());
PQclear(result);
return code;
}
PQclear(result);
return ADBC_STATUS_OK;
}
AdbcStatusCode Execute(PGconn* conn, int64_t* rows_affected, struct AdbcError* error) {
if (rows_affected) *rows_affected = 0;
PGresult* result = nullptr;
while (true) {
Handle<struct ArrowArray> array;
int res = bind->get_next(&bind.value, &array.value);
if (res != 0) {
SetError(error,
"[libpq] Failed to read next batch from stream of bind parameters: "
"(%d) %s %s",
res, std::strerror(res), bind->get_last_error(&bind.value));
return ADBC_STATUS_IO;
}
if (!array->release) break;
Handle<struct ArrowArrayView> array_view;
// TODO: include error messages
CHECK_NA(
INTERNAL,
ArrowArrayViewInitFromSchema(&array_view.value, &bind_schema.value, nullptr),
error);
CHECK_NA(INTERNAL, ArrowArrayViewSetArray(&array_view.value, &array.value, nullptr),
error);
for (int64_t row = 0; row < array->length; row++) {
for (int64_t col = 0; col < array_view->n_children; col++) {
if (ArrowArrayViewIsNull(array_view->children[col], row)) {
param_values[col] = nullptr;
continue;
} else {
param_values[col] = param_values_buffer.data() + param_values_offsets[col];
}
switch (bind_schema_fields[col].type) {
case ArrowType::NANOARROW_TYPE_BOOL: {
const int8_t val = ArrowBitGet(
array_view->children[col]->buffer_views[1].data.as_uint8, row);
std::memcpy(param_values[col], &val, sizeof(int8_t));
break;
}
case ArrowType::NANOARROW_TYPE_INT8: {
const int16_t val =
array_view->children[col]->buffer_views[1].data.as_int8[row];
const uint16_t value = ToNetworkInt16(val);
std::memcpy(param_values[col], &value, sizeof(int16_t));
break;
}
case ArrowType::NANOARROW_TYPE_INT16: {
const uint16_t value = ToNetworkInt16(
array_view->children[col]->buffer_views[1].data.as_int16[row]);
std::memcpy(param_values[col], &value, sizeof(int16_t));
break;
}
case ArrowType::NANOARROW_TYPE_INT32: {
const uint32_t value = ToNetworkInt32(
array_view->children[col]->buffer_views[1].data.as_int32[row]);
std::memcpy(param_values[col], &value, sizeof(int32_t));
break;
}
case ArrowType::NANOARROW_TYPE_INT64: {
const int64_t value = ToNetworkInt64(
array_view->children[col]->buffer_views[1].data.as_int64[row]);
std::memcpy(param_values[col], &value, sizeof(int64_t));
break;
}
case ArrowType::NANOARROW_TYPE_FLOAT: {
const uint32_t value = ToNetworkFloat4(
array_view->children[col]->buffer_views[1].data.as_float[row]);
std::memcpy(param_values[col], &value, sizeof(uint32_t));
break;
}
case ArrowType::NANOARROW_TYPE_DOUBLE: {
const uint64_t value = ToNetworkFloat8(
array_view->children[col]->buffer_views[1].data.as_double[row]);
std::memcpy(param_values[col], &value, sizeof(uint64_t));
break;
}
case ArrowType::NANOARROW_TYPE_STRING:
case ArrowType::NANOARROW_TYPE_LARGE_STRING:
case ArrowType::NANOARROW_TYPE_BINARY: {
const ArrowBufferView view =
ArrowArrayViewGetBytesUnsafe(array_view->children[col], row);
// TODO: overflow check?
param_lengths[col] = static_cast<int>(view.size_bytes);
param_values[col] = const_cast<char*>(view.data.as_char);
break;
}
case ArrowType::NANOARROW_TYPE_DATE32: {
// 2000-01-01
constexpr int32_t kPostgresDateEpoch = 10957;
const int32_t raw_value =
array_view->children[col]->buffer_views[1].data.as_int32[row];
if (raw_value < INT32_MIN + kPostgresDateEpoch) {
SetError(error, "[libpq] Field #%" PRId64 "%s%s%s%" PRId64 "%s", col + 1,
"('", bind_schema->children[col]->name, "') Row #", row + 1,
"has value which exceeds postgres date limits");
return ADBC_STATUS_INVALID_ARGUMENT;
}
const uint32_t value = ToNetworkInt32(raw_value - kPostgresDateEpoch);
std::memcpy(param_values[col], &value, sizeof(int32_t));
break;
}
case ArrowType::NANOARROW_TYPE_DURATION:
case ArrowType::NANOARROW_TYPE_TIMESTAMP: {
int64_t val = array_view->children[col]->buffer_views[1].data.as_int64[row];
// 2000-01-01 00:00:00.000000 in microseconds
constexpr int64_t kPostgresTimestampEpoch = 946684800000000;
bool overflow_safe = true;
auto unit = bind_schema_fields[col].time_unit;
switch (unit) {
case NANOARROW_TIME_UNIT_SECOND:
if ((overflow_safe = val <= kMaxSafeSecondsToMicros &&
val >= kMinSafeSecondsToMicros)) {
val *= 1000000;
}
break;
case NANOARROW_TIME_UNIT_MILLI:
if ((overflow_safe = val <= kMaxSafeMillisToMicros &&
val >= kMinSafeMillisToMicros)) {
val *= 1000;
}
break;
case NANOARROW_TIME_UNIT_MICRO:
break;
case NANOARROW_TIME_UNIT_NANO:
val /= 1000;
break;
}
if (!overflow_safe) {
SetError(error,
"[libpq] Field #%" PRId64 " ('%s') Row #%" PRId64
" has value '%" PRIi64
"' which exceeds PostgreSQL timestamp limits",
col + 1, bind_schema->children[col]->name, row + 1,
array_view->children[col]->buffer_views[1].data.as_int64[row]);
return ADBC_STATUS_INVALID_ARGUMENT;
}
if (bind_schema_fields[col].type == ArrowType::NANOARROW_TYPE_TIMESTAMP) {
const uint64_t value = ToNetworkInt64(val - kPostgresTimestampEpoch);
std::memcpy(param_values[col], &value, sizeof(int64_t));
} else if (bind_schema_fields[col].type ==
ArrowType::NANOARROW_TYPE_DURATION) {
// postgres stores an interval as a 64 bit offset in microsecond
// resolution alongside a 32 bit day and 32 bit month
// for now we just send 0 for the day / month values
const uint64_t value = ToNetworkInt64(val);
std::memcpy(param_values[col], &value, sizeof(int64_t));
std::memset(param_values[col] + sizeof(int64_t), 0, sizeof(int64_t));
}
break;
}
case ArrowType::NANOARROW_TYPE_INTERVAL_MONTH_DAY_NANO: {
struct ArrowInterval interval;
ArrowIntervalInit(&interval, NANOARROW_TYPE_INTERVAL_MONTH_DAY_NANO);
ArrowArrayViewGetIntervalUnsafe(array_view->children[col], row, &interval);
const uint32_t months = ToNetworkInt32(interval.months);
const uint32_t days = ToNetworkInt32(interval.days);
const uint64_t ms = ToNetworkInt64(interval.ns / 1000);
std::memcpy(param_values[col], &ms, sizeof(uint64_t));
std::memcpy(param_values[col] + sizeof(uint64_t), &days, sizeof(uint32_t));
std::memcpy(param_values[col] + sizeof(uint64_t) + sizeof(uint32_t),
&months, sizeof(uint32_t));
break;
}
default:
SetError(error, "%s%" PRId64 "%s%s%s%s", "[libpq] Field #", col + 1, " ('",
bind_schema->children[col]->name,
"') has unsupported type for ingestion ",
ArrowTypeString(bind_schema_fields[col].type));
return ADBC_STATUS_NOT_IMPLEMENTED;
}
}
result = PQexecPrepared(conn, /*stmtName=*/"",
/*nParams=*/bind_schema->n_children, param_values.data(),
param_lengths.data(), param_formats.data(),
/*resultFormat=*/0 /*text*/);
ExecStatusType pg_status = PQresultStatus(result);
if (pg_status != PGRES_COMMAND_OK) {
AdbcStatusCode code = SetError(
error, result, "[libpq] Failed to execute prepared statement: %s %s",
PQresStatus(pg_status), PQerrorMessage(conn));
PQclear(result);
return code;
}
PQclear(result);
}
if (rows_affected) *rows_affected += array->length;
if (has_tz_field) {
std::string reset_query = "SET TIME ZONE '" + tz_setting + "'";
PGresult* reset_tz_result = PQexec(conn, reset_query.c_str());
if (PQresultStatus(reset_tz_result) != PGRES_COMMAND_OK) {
AdbcStatusCode code =
SetError(error, reset_tz_result, "[libpq] Failed to reset time zone: %s",
PQerrorMessage(conn));
PQclear(reset_tz_result);
return code;
}
PQclear(reset_tz_result);
PGresult* commit_result = PQexec(conn, "COMMIT");
if (PQresultStatus(commit_result) != PGRES_COMMAND_OK) {
AdbcStatusCode code =
SetError(error, commit_result, "[libpq] Failed to commit transaction: %s",
PQerrorMessage(conn));
PQclear(commit_result);
return code;
}
PQclear(commit_result);
}
}
return ADBC_STATUS_OK;
}
};
} // namespace
int TupleReader::GetSchema(struct ArrowSchema* out) {
assert(copy_reader_ != nullptr);
int na_res = copy_reader_->GetSchema(out);
if (out->release == nullptr) {
SetError(&error_, "[libpq] Result set was already consumed or freed");
status_ = ADBC_STATUS_INVALID_STATE;
return AdbcStatusCodeToErrno(status_);
} else if (na_res != NANOARROW_OK) {
// e.g., Can't allocate memory
SetError(&error_, "[libpq] Error copying schema");
status_ = ADBC_STATUS_INTERNAL;
}
return na_res;
}
int TupleReader::InitQueryAndFetchFirst(struct ArrowError* error) {
// Fetch + parse the header
int get_copy_res = PQgetCopyData(conn_, &pgbuf_, /*async=*/0);
data_.size_bytes = get_copy_res;
data_.data.as_char = pgbuf_;
if (get_copy_res == -2) {
SetError(&error_, "[libpq] Fetch header failed: %s", PQerrorMessage(conn_));
status_ = ADBC_STATUS_IO;
return AdbcStatusCodeToErrno(status_);
}
int na_res = copy_reader_->ReadHeader(&data_, error);
if (na_res != NANOARROW_OK) {
SetError(&error_, "[libpq] ReadHeader failed: %s", error->message);
status_ = ADBC_STATUS_IO;
return AdbcStatusCodeToErrno(status_);
}
return NANOARROW_OK;
}
int TupleReader::AppendRowAndFetchNext(struct ArrowError* error) {
// Parse the result (the header AND the first row are included in the first
// call to PQgetCopyData())
int na_res = copy_reader_->ReadRecord(&data_, error);
if (na_res != NANOARROW_OK && na_res != ENODATA) {
SetError(&error_, "[libpq] ReadRecord failed at row %" PRId64 ": %s", row_id_,
error->message);
status_ = ADBC_STATUS_IO;
return na_res;
}
row_id_++;
// Fetch + check
PQfreemem(pgbuf_);
pgbuf_ = nullptr;
int get_copy_res = PQgetCopyData(conn_, &pgbuf_, /*async=*/0);
data_.size_bytes = get_copy_res;
data_.data.as_char = pgbuf_;
if (get_copy_res == -2) {
SetError(&error_, "[libpq] PQgetCopyData failed at row %" PRId64 ": %s", row_id_,
PQerrorMessage(conn_));
status_ = ADBC_STATUS_IO;
return AdbcStatusCodeToErrno(status_);
} else if (get_copy_res == -1) {
// Returned when COPY has finished successfully
return ENODATA;
} else if ((copy_reader_->array_size_approx_bytes() + get_copy_res) >=
batch_size_hint_bytes_) {
// Appending the next row will result in an array larger than requested.
// Return EOVERFLOW to force GetNext() to build the current result and return.
return EOVERFLOW;
} else {
return NANOARROW_OK;
}
}
int TupleReader::BuildOutput(struct ArrowArray* out, struct ArrowError* error) {
if (copy_reader_->array_size_approx_bytes() == 0) {
out->release = nullptr;
return NANOARROW_OK;
}
int na_res = copy_reader_->GetArray(out, error);
if (na_res != NANOARROW_OK) {
SetError(&error_, "[libpq] Failed to build result array: %s", error->message);
status_ = ADBC_STATUS_INTERNAL;
return na_res;
}
return NANOARROW_OK;
}
int TupleReader::GetNext(struct ArrowArray* out) {
if (is_finished_) {
out->release = nullptr;
return 0;
}
struct ArrowError error;
error.message[0] = '\0';
if (row_id_ == -1) {
NANOARROW_RETURN_NOT_OK(InitQueryAndFetchFirst(&error));
row_id_++;
}
int na_res;
do {
na_res = AppendRowAndFetchNext(&error);
if (na_res == EOVERFLOW) {
// The result would be too big to return if we appended the row. When EOVERFLOW is
// returned, the copy reader leaves the output in a valid state. The data is left in
// pg_buf_/data_ and will attempt to be appended on the next call to GetNext()
return BuildOutput(out, &error);
}
} while (na_res == NANOARROW_OK);
if (na_res != ENODATA) {
return na_res;
}
is_finished_ = true;
// Finish the result properly and return the last result. Note that BuildOutput() may
// set tmp.release = nullptr if there were zero rows in the copy reader (can
// occur in an overflow scenario).
struct ArrowArray tmp;
NANOARROW_RETURN_NOT_OK(BuildOutput(&tmp, &error));
PQclear(result_);
// Check the server-side response
result_ = PQgetResult(conn_);
const ExecStatusType pq_status = PQresultStatus(result_);
if (pq_status != PGRES_COMMAND_OK) {
const char* sqlstate = PQresultErrorField(result_, PG_DIAG_SQLSTATE);
SetError(&error_, result_, "[libpq] Query failed [%s]: %s", PQresStatus(pq_status),
PQresultErrorMessage(result_));
if (tmp.release != nullptr) {
tmp.release(&tmp);
}
if (sqlstate != nullptr && std::strcmp(sqlstate, "57014") == 0) {
status_ = ADBC_STATUS_CANCELLED;
} else {
status_ = ADBC_STATUS_IO;
}
return AdbcStatusCodeToErrno(status_);
}
ArrowArrayMove(&tmp, out);
return NANOARROW_OK;
}
void TupleReader::Release() {
if (error_.release) {
error_.release(&error_);
}
error_ = ADBC_ERROR_INIT;
status_ = ADBC_STATUS_OK;
if (result_) {
PQclear(result_);
result_ = nullptr;
}
if (pgbuf_) {
PQfreemem(pgbuf_);
pgbuf_ = nullptr;
}
if (copy_reader_) {
copy_reader_.reset();
}
is_finished_ = false;
row_id_ = -1;
}
void TupleReader::ExportTo(struct ArrowArrayStream* stream) {
stream->get_schema = &GetSchemaTrampoline;
stream->get_next = &GetNextTrampoline;
stream->get_last_error = &GetLastErrorTrampoline;
stream->release = &ReleaseTrampoline;
stream->private_data = this;
}
const struct AdbcError* TupleReader::ErrorFromArrayStream(struct ArrowArrayStream* stream,
AdbcStatusCode* status) {
if (!stream->private_data || stream->release != &ReleaseTrampoline) {
return nullptr;
}
TupleReader* reader = static_cast<TupleReader*>(stream->private_data);
if (status) {
*status = reader->status_;
}
return &reader->error_;
}
int TupleReader::GetSchemaTrampoline(struct ArrowArrayStream* self,
struct ArrowSchema* out) {
if (!self || !self->private_data) return EINVAL;
TupleReader* reader = static_cast<TupleReader*>(self->private_data);
return reader->GetSchema(out);
}
int TupleReader::GetNextTrampoline(struct ArrowArrayStream* self,
struct ArrowArray* out) {
if (!self || !self->private_data) return EINVAL;
TupleReader* reader = static_cast<TupleReader*>(self->private_data);
return reader->GetNext(out);
}
const char* TupleReader::GetLastErrorTrampoline(struct ArrowArrayStream* self) {
if (!self || !self->private_data) return nullptr;
TupleReader* reader = static_cast<TupleReader*>(self->private_data);
return reader->last_error();
}
void TupleReader::ReleaseTrampoline(struct ArrowArrayStream* self) {
if (!self || !self->private_data) return;
TupleReader* reader = static_cast<TupleReader*>(self->private_data);
reader->Release();
self->private_data = nullptr;
self->release = nullptr;
}
AdbcStatusCode PostgresStatement::New(struct AdbcConnection* connection,
struct AdbcError* error) {
if (!connection || !connection->private_data) {
SetError(error, "%s", "[libpq] Must provide an initialized AdbcConnection");
return ADBC_STATUS_INVALID_ARGUMENT;
}
connection_ =
*reinterpret_cast<std::shared_ptr<PostgresConnection>*>(connection->private_data);
type_resolver_ = connection_->type_resolver();
reader_.conn_ = connection_->conn();
return ADBC_STATUS_OK;
}
AdbcStatusCode PostgresStatement::Bind(struct ArrowArray* values,
struct ArrowSchema* schema,
struct AdbcError* error) {
if (!values || !values->release) {
SetError(error, "%s", "[libpq] Must provide non-NULL array");
return ADBC_STATUS_INVALID_ARGUMENT;
} else if (!schema || !schema->release) {
SetError(error, "%s", "[libpq] Must provide non-NULL schema");
return ADBC_STATUS_INVALID_ARGUMENT;
}
if (bind_.release) bind_.release(&bind_);
// Make a one-value stream
bind_.private_data = new OneValueStream{*schema, *values};
bind_.get_schema = &OneValueStream::GetSchema;
bind_.get_next = &OneValueStream::GetNext;
bind_.get_last_error = &OneValueStream::GetLastError;
bind_.release = &OneValueStream::Release;
std::memset(values, 0, sizeof(*values));
std::memset(schema, 0, sizeof(*schema));
return ADBC_STATUS_OK;
}
AdbcStatusCode PostgresStatement::Bind(struct ArrowArrayStream* stream,
struct AdbcError* error) {
if (!stream || !stream->release) {
SetError(error, "%s", "[libpq] Must provide non-NULL stream");
return ADBC_STATUS_INVALID_ARGUMENT;
}
// Move stream
if (bind_.release) bind_.release(&bind_);
bind_ = *stream;
std::memset(stream, 0, sizeof(*stream));
return ADBC_STATUS_OK;
}
AdbcStatusCode PostgresStatement::Cancel(struct AdbcError* error) {
// Ultimately the same underlying PGconn
return connection_->Cancel(error);
}
AdbcStatusCode PostgresStatement::CreateBulkTable(
const std::string& current_schema, const struct ArrowSchema& source_schema,
const std::vector<struct ArrowSchemaView>& source_schema_fields,
std::string* escaped_table, struct AdbcError* error) {
PGconn* conn = connection_->conn();
if (!ingest_.db_schema.empty() && ingest_.temporary) {
SetError(error, "[libpq] Cannot set both %s and %s",
ADBC_INGEST_OPTION_TARGET_DB_SCHEMA, ADBC_INGEST_OPTION_TEMPORARY);
return ADBC_STATUS_INVALID_STATE;
}
{
if (!ingest_.db_schema.empty()) {
char* escaped =
PQescapeIdentifier(conn, ingest_.db_schema.c_str(), ingest_.db_schema.size());
if (escaped == nullptr) {
SetError(error, "[libpq] Failed to escape target schema %s for ingestion: %s",
ingest_.db_schema.c_str(), PQerrorMessage(conn));
return ADBC_STATUS_INTERNAL;
}
*escaped_table += escaped;
*escaped_table += " . ";
PQfreemem(escaped);
} else if (ingest_.temporary) {
// OK to be redundant (CREATE TEMPORARY TABLE pg_temp.foo)
*escaped_table += "pg_temp . ";
} else {
// Explicitly specify the current schema to avoid any temporary tables
// shadowing this table
char* escaped =
PQescapeIdentifier(conn, current_schema.c_str(), current_schema.size());
*escaped_table += escaped;
*escaped_table += " . ";
PQfreemem(escaped);
}
if (!ingest_.target.empty()) {
char* escaped =
PQescapeIdentifier(conn, ingest_.target.c_str(), ingest_.target.size());
if (escaped == nullptr) {
SetError(error, "[libpq] Failed to escape target table %s for ingestion: %s",
ingest_.target.c_str(), PQerrorMessage(conn));
return ADBC_STATUS_INTERNAL;
}
*escaped_table += escaped;
PQfreemem(escaped);
}
}
std::string create;
if (ingest_.temporary) {
create = "CREATE TEMPORARY TABLE ";
} else {
create = "CREATE TABLE ";
}
switch (ingest_.mode) {
case IngestMode::kCreate:
// Nothing to do
break;
case IngestMode::kAppend:
return ADBC_STATUS_OK;
case IngestMode::kReplace: {
std::string drop = "DROP TABLE IF EXISTS " + *escaped_table;
PGresult* result = PQexecParams(conn, drop.c_str(), /*nParams=*/0,
/*paramTypes=*/nullptr, /*paramValues=*/nullptr,
/*paramLengths=*/nullptr, /*paramFormats=*/nullptr,
/*resultFormat=*/1 /*(binary)*/);
if (PQresultStatus(result) != PGRES_COMMAND_OK) {
AdbcStatusCode code =
SetError(error, result, "[libpq] Failed to drop table: %s\nQuery was: %s",
PQerrorMessage(conn), drop.c_str());
PQclear(result);
return code;
}
PQclear(result);
break;
}
case IngestMode::kCreateAppend:
create += "IF NOT EXISTS ";
break;
}
create += *escaped_table;
create += " (";
for (size_t i = 0; i < source_schema_fields.size(); i++) {
if (i > 0) create += ", ";
const char* unescaped = source_schema.children[i]->name;
char* escaped = PQescapeIdentifier(conn, unescaped, std::strlen(unescaped));
if (escaped == nullptr) {
SetError(error, "[libpq] Failed to escape column %s for ingestion: %s", unescaped,
PQerrorMessage(conn));
return ADBC_STATUS_INTERNAL;
}
create += escaped;
PQfreemem(escaped);
switch (source_schema_fields[i].type) {
case ArrowType::NANOARROW_TYPE_BOOL:
create += " BOOLEAN";
break;
case ArrowType::NANOARROW_TYPE_INT8:
case ArrowType::NANOARROW_TYPE_INT16:
create += " SMALLINT";
break;
case ArrowType::NANOARROW_TYPE_INT32:
create += " INTEGER";
break;
case ArrowType::NANOARROW_TYPE_INT64:
create += " BIGINT";
break;
case ArrowType::NANOARROW_TYPE_FLOAT:
create += " REAL";
break;
case ArrowType::NANOARROW_TYPE_DOUBLE:
create += " DOUBLE PRECISION";
break;
case ArrowType::NANOARROW_TYPE_STRING:
case ArrowType::NANOARROW_TYPE_LARGE_STRING:
create += " TEXT";
break;
case ArrowType::NANOARROW_TYPE_BINARY:
create += " BYTEA";
break;
case ArrowType::NANOARROW_TYPE_DATE32:
create += " DATE";
break;
case ArrowType::NANOARROW_TYPE_TIMESTAMP:
if (strcmp("", source_schema_fields[i].timezone)) {
create += " TIMESTAMPTZ";
} else {
create += " TIMESTAMP";
}
break;
case ArrowType::NANOARROW_TYPE_DURATION:
case ArrowType::NANOARROW_TYPE_INTERVAL_MONTH_DAY_NANO:
create += " INTERVAL";
break;
default:
SetError(error, "%s%" PRIu64 "%s%s%s%s", "[libpq] Field #",
static_cast<uint64_t>(i + 1), " ('", source_schema.children[i]->name,
"') has unsupported type for ingestion ",
ArrowTypeString(source_schema_fields[i].type));
return ADBC_STATUS_NOT_IMPLEMENTED;
}
}
create += ")";
SetError(error, "%s%s", "[libpq] ", create.c_str());
PGresult* result = PQexecParams(conn, create.c_str(), /*nParams=*/0,
/*paramTypes=*/nullptr, /*paramValues=*/nullptr,