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
|
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#ifndef ROCKSDB_LITE
#include "table/plain_table_reader.h"
#include <string>
#include <vector>
#include "db/dbformat.h"
#include "rocksdb/cache.h"
#include "rocksdb/comparator.h"
#include "rocksdb/env.h"
#include "rocksdb/filter_policy.h"
#include "rocksdb/options.h"
#include "rocksdb/statistics.h"
#include "table/block.h"
#include "table/bloom_block.h"
#include "table/filter_block.h"
#include "table/format.h"
#include "table/internal_iterator.h"
#include "table/meta_blocks.h"
#include "table/two_level_iterator.h"
#include "table/plain_table_factory.h"
#include "table/plain_table_key_coding.h"
#include "table/get_context.h"
#include "monitoring/histogram.h"
#include "monitoring/perf_context_imp.h"
#include "util/arena.h"
#include "util/coding.h"
#include "util/dynamic_bloom.h"
#include "util/hash.h"
#include "util/murmurhash.h"
#include "util/stop_watch.h"
#include "util/string_util.h"
namespace rocksdb {
namespace {
// Safely getting a uint32_t element from a char array, where, starting from
// `base`, every 4 bytes are considered as an fixed 32 bit integer.
inline uint32_t GetFixed32Element(const char* base, size_t offset) {
return DecodeFixed32(base + offset * sizeof(uint32_t));
}
} // namespace
// Iterator to iterate IndexedTable
class PlainTableIterator : public InternalIterator {
public:
explicit PlainTableIterator(PlainTableReader* table, bool use_prefix_seek);
~PlainTableIterator() override;
bool Valid() const override;
void SeekToFirst() override;
void SeekToLast() override;
void Seek(const Slice& target) override;
void SeekForPrev(const Slice& target) override;
void Next() override;
void Prev() override;
Slice key() const override;
Slice value() const override;
Status status() const override;
private:
PlainTableReader* table_;
PlainTableKeyDecoder decoder_;
bool use_prefix_seek_;
uint32_t offset_;
uint32_t next_offset_;
Slice key_;
Slice value_;
Status status_;
// No copying allowed
PlainTableIterator(const PlainTableIterator&) = delete;
void operator=(const Iterator&) = delete;
};
extern const uint64_t kPlainTableMagicNumber;
PlainTableReader::PlainTableReader(
const ImmutableCFOptions& ioptions,
std::unique_ptr<RandomAccessFileReader>&& file,
const EnvOptions& storage_options, const InternalKeyComparator& icomparator,
EncodingType encoding_type, uint64_t file_size,
const TableProperties* table_properties,
const SliceTransform* prefix_extractor)
: internal_comparator_(icomparator),
encoding_type_(encoding_type),
full_scan_mode_(false),
user_key_len_(static_cast<uint32_t>(table_properties->fixed_key_len)),
prefix_extractor_(prefix_extractor),
enable_bloom_(false),
bloom_(6),
file_info_(std::move(file), storage_options,
static_cast<uint32_t>(table_properties->data_size)),
ioptions_(ioptions),
file_size_(file_size),
table_properties_(nullptr) {}
PlainTableReader::~PlainTableReader() {
}
Status PlainTableReader::Open(
const ImmutableCFOptions& ioptions, const EnvOptions& env_options,
const InternalKeyComparator& internal_comparator,
std::unique_ptr<RandomAccessFileReader>&& file, uint64_t file_size,
std::unique_ptr<TableReader>* table_reader, const int bloom_bits_per_key,
double hash_table_ratio, size_t index_sparseness, size_t huge_page_tlb_size,
bool full_scan_mode, const bool immortal_table,
const SliceTransform* prefix_extractor) {
if (file_size > PlainTableIndex::kMaxFileSize) {
return Status::NotSupported("File is too large for PlainTableReader!");
}
TableProperties* props = nullptr;
auto s = ReadTableProperties(file.get(), file_size, kPlainTableMagicNumber,
ioptions, &props,
true /* compression_type_missing */);
if (!s.ok()) {
return s;
}
assert(hash_table_ratio >= 0.0);
auto& user_props = props->user_collected_properties;
auto prefix_extractor_in_file = props->prefix_extractor_name;
if (!full_scan_mode &&
!prefix_extractor_in_file.empty() /* old version sst file*/
&& prefix_extractor_in_file != "nullptr") {
if (!prefix_extractor) {
return Status::InvalidArgument(
"Prefix extractor is missing when opening a PlainTable built "
"using a prefix extractor");
} else if (prefix_extractor_in_file.compare(prefix_extractor->Name()) !=
0) {
return Status::InvalidArgument(
"Prefix extractor given doesn't match the one used to build "
"PlainTable");
}
}
EncodingType encoding_type = kPlain;
auto encoding_type_prop =
user_props.find(PlainTablePropertyNames::kEncodingType);
if (encoding_type_prop != user_props.end()) {
encoding_type = static_cast<EncodingType>(
DecodeFixed32(encoding_type_prop->second.c_str()));
}
std::unique_ptr<PlainTableReader> new_reader(new PlainTableReader(
ioptions, std::move(file), env_options, internal_comparator,
encoding_type, file_size, props, prefix_extractor));
s = new_reader->MmapDataIfNeeded();
if (!s.ok()) {
return s;
}
if (!full_scan_mode) {
s = new_reader->PopulateIndex(props, bloom_bits_per_key, hash_table_ratio,
index_sparseness, huge_page_tlb_size);
if (!s.ok()) {
return s;
}
} else {
// Flag to indicate it is a full scan mode so that none of the indexes
// can be used.
new_reader->full_scan_mode_ = true;
}
if (immortal_table && new_reader->file_info_.is_mmap_mode) {
new_reader->dummy_cleanable_.reset(new Cleanable());
}
*table_reader = std::move(new_reader);
return s;
}
void PlainTableReader::SetupForCompaction() {
}
InternalIterator* PlainTableReader::NewIterator(
const ReadOptions& options, const SliceTransform* /* prefix_extractor */,
Arena* arena, bool /*skip_filters*/, bool /*for_compaction*/) {
bool use_prefix_seek = !IsTotalOrderMode() && !options.total_order_seek;
if (arena == nullptr) {
return new PlainTableIterator(this, use_prefix_seek);
} else {
auto mem = arena->AllocateAligned(sizeof(PlainTableIterator));
return new (mem) PlainTableIterator(this, use_prefix_seek);
}
}
Status PlainTableReader::PopulateIndexRecordList(
PlainTableIndexBuilder* index_builder,
std::vector<uint32_t>* prefix_hashes) {
Slice prev_key_prefix_slice;
std::string prev_key_prefix_buf;
uint32_t pos = data_start_offset_;
bool is_first_record = true;
Slice key_prefix_slice;
PlainTableKeyDecoder decoder(&file_info_, encoding_type_, user_key_len_,
prefix_extractor_);
while (pos < file_info_.data_end_offset) {
uint32_t key_offset = pos;
ParsedInternalKey key;
Slice value_slice;
bool seekable = false;
Status s = Next(&decoder, &pos, &key, nullptr, &value_slice, &seekable);
if (!s.ok()) {
return s;
}
key_prefix_slice = GetPrefix(key);
if (enable_bloom_) {
bloom_.AddHash(GetSliceHash(key.user_key));
} else {
if (is_first_record || prev_key_prefix_slice != key_prefix_slice) {
if (!is_first_record) {
prefix_hashes->push_back(GetSliceHash(prev_key_prefix_slice));
}
if (file_info_.is_mmap_mode) {
prev_key_prefix_slice = key_prefix_slice;
} else {
prev_key_prefix_buf = key_prefix_slice.ToString();
prev_key_prefix_slice = prev_key_prefix_buf;
}
}
}
index_builder->AddKeyPrefix(GetPrefix(key), key_offset);
if (!seekable && is_first_record) {
return Status::Corruption("Key for a prefix is not seekable");
}
is_first_record = false;
}
prefix_hashes->push_back(GetSliceHash(key_prefix_slice));
auto s = index_.InitFromRawData(index_builder->Finish());
return s;
}
void PlainTableReader::AllocateAndFillBloom(
int bloom_bits_per_key, int num_prefixes, size_t huge_page_tlb_size,
std::vector<uint32_t>* prefix_hashes) {
if (!IsTotalOrderMode()) {
uint32_t bloom_total_bits = num_prefixes * bloom_bits_per_key;
if (bloom_total_bits > 0) {
enable_bloom_ = true;
bloom_.SetTotalBits(&arena_, bloom_total_bits, ioptions_.bloom_locality,
huge_page_tlb_size, ioptions_.info_log);
FillBloom(prefix_hashes);
}
}
}
void PlainTableReader::FillBloom(std::vector<uint32_t>* prefix_hashes) {
assert(bloom_.IsInitialized());
for (auto prefix_hash : *prefix_hashes) {
bloom_.AddHash(prefix_hash);
}
}
Status PlainTableReader::MmapDataIfNeeded() {
if (file_info_.is_mmap_mode) {
// Get mmapped memory.
return file_info_.file->Read(0, static_cast<size_t>(file_size_), &file_info_.file_data, nullptr);
}
return Status::OK();
}
Status PlainTableReader::PopulateIndex(TableProperties* props,
int bloom_bits_per_key,
double hash_table_ratio,
size_t index_sparseness,
size_t huge_page_tlb_size) {
assert(props != nullptr);
table_properties_.reset(props);
BlockContents index_block_contents;
Status s = ReadMetaBlock(file_info_.file.get(), nullptr /* prefetch_buffer */,
file_size_, kPlainTableMagicNumber, ioptions_,
PlainTableIndexBuilder::kPlainTableIndexBlock,
&index_block_contents,
true /* compression_type_missing */);
bool index_in_file = s.ok();
BlockContents bloom_block_contents;
bool bloom_in_file = false;
// We only need to read the bloom block if index block is in file.
if (index_in_file) {
s = ReadMetaBlock(file_info_.file.get(), nullptr /* prefetch_buffer */,
file_size_, kPlainTableMagicNumber, ioptions_,
BloomBlockBuilder::kBloomBlock, &bloom_block_contents,
true /* compression_type_missing */);
bloom_in_file = s.ok() && bloom_block_contents.data.size() > 0;
}
Slice* bloom_block;
if (bloom_in_file) {
// If bloom_block_contents.allocation is not empty (which will be the case
// for non-mmap mode), it holds the alloated memory for the bloom block.
// It needs to be kept alive to keep `bloom_block` valid.
bloom_block_alloc_ = std::move(bloom_block_contents.allocation);
bloom_block = &bloom_block_contents.data;
} else {
bloom_block = nullptr;
}
Slice* index_block;
if (index_in_file) {
// If index_block_contents.allocation is not empty (which will be the case
// for non-mmap mode), it holds the alloated memory for the index block.
// It needs to be kept alive to keep `index_block` valid.
index_block_alloc_ = std::move(index_block_contents.allocation);
index_block = &index_block_contents.data;
} else {
index_block = nullptr;
}
if ((prefix_extractor_ == nullptr) && (hash_table_ratio != 0)) {
// moptions.prefix_extractor is requried for a hash-based look-up.
return Status::NotSupported(
"PlainTable requires a prefix extractor enable prefix hash mode.");
}
// First, read the whole file, for every kIndexIntervalForSamePrefixKeys rows
// for a prefix (starting from the first one), generate a record of (hash,
// offset) and append it to IndexRecordList, which is a data structure created
// to store them.
if (!index_in_file) {
// Allocate bloom filter here for total order mode.
if (IsTotalOrderMode()) {
uint32_t num_bloom_bits =
static_cast<uint32_t>(table_properties_->num_entries) *
bloom_bits_per_key;
if (num_bloom_bits > 0) {
enable_bloom_ = true;
bloom_.SetTotalBits(&arena_, num_bloom_bits, ioptions_.bloom_locality,
huge_page_tlb_size, ioptions_.info_log);
}
}
} else if (bloom_in_file) {
enable_bloom_ = true;
auto num_blocks_property = props->user_collected_properties.find(
PlainTablePropertyNames::kNumBloomBlocks);
uint32_t num_blocks = 0;
if (num_blocks_property != props->user_collected_properties.end()) {
Slice temp_slice(num_blocks_property->second);
if (!GetVarint32(&temp_slice, &num_blocks)) {
num_blocks = 0;
}
}
// cast away const qualifier, because bloom_ won't be changed
bloom_.SetRawData(
const_cast<unsigned char*>(
reinterpret_cast<const unsigned char*>(bloom_block->data())),
static_cast<uint32_t>(bloom_block->size()) * 8, num_blocks);
} else {
// Index in file but no bloom in file. Disable bloom filter in this case.
enable_bloom_ = false;
bloom_bits_per_key = 0;
}
PlainTableIndexBuilder index_builder(&arena_, ioptions_, prefix_extractor_,
index_sparseness, hash_table_ratio,
huge_page_tlb_size);
std::vector<uint32_t> prefix_hashes;
if (!index_in_file) {
s = PopulateIndexRecordList(&index_builder, &prefix_hashes);
if (!s.ok()) {
return s;
}
} else {
s = index_.InitFromRawData(*index_block);
if (!s.ok()) {
return s;
}
}
if (!index_in_file) {
// Calculated bloom filter size and allocate memory for
// bloom filter based on the number of prefixes, then fill it.
AllocateAndFillBloom(bloom_bits_per_key, index_.GetNumPrefixes(),
huge_page_tlb_size, &prefix_hashes);
}
// Fill two table properties.
if (!index_in_file) {
props->user_collected_properties["plain_table_hash_table_size"] =
ToString(index_.GetIndexSize() * PlainTableIndex::kOffsetLen);
props->user_collected_properties["plain_table_sub_index_size"] =
ToString(index_.GetSubIndexSize());
} else {
props->user_collected_properties["plain_table_hash_table_size"] =
ToString(0);
props->user_collected_properties["plain_table_sub_index_size"] =
ToString(0);
}
return Status::OK();
}
Status PlainTableReader::GetOffset(PlainTableKeyDecoder* decoder,
const Slice& target, const Slice& prefix,
uint32_t prefix_hash, bool& prefix_matched,
uint32_t* offset) const {
prefix_matched = false;
uint32_t prefix_index_offset;
auto res = index_.GetOffset(prefix_hash, &prefix_index_offset);
if (res == PlainTableIndex::kNoPrefixForBucket) {
*offset = file_info_.data_end_offset;
return Status::OK();
} else if (res == PlainTableIndex::kDirectToFile) {
*offset = prefix_index_offset;
return Status::OK();
}
// point to sub-index, need to do a binary search
uint32_t upper_bound;
const char* base_ptr =
index_.GetSubIndexBasePtrAndUpperBound(prefix_index_offset, &upper_bound);
uint32_t low = 0;
uint32_t high = upper_bound;
ParsedInternalKey mid_key;
ParsedInternalKey parsed_target;
if (!ParseInternalKey(target, &parsed_target)) {
return Status::Corruption(Slice());
}
// The key is between [low, high). Do a binary search between it.
while (high - low > 1) {
uint32_t mid = (high + low) / 2;
uint32_t file_offset = GetFixed32Element(base_ptr, mid);
uint32_t tmp;
Status s = decoder->NextKeyNoValue(file_offset, &mid_key, nullptr, &tmp);
if (!s.ok()) {
return s;
}
int cmp_result = internal_comparator_.Compare(mid_key, parsed_target);
if (cmp_result < 0) {
low = mid;
} else {
if (cmp_result == 0) {
// Happen to have found the exact key or target is smaller than the
// first key after base_offset.
prefix_matched = true;
*offset = file_offset;
return Status::OK();
} else {
high = mid;
}
}
}
// Both of the key at the position low or low+1 could share the same
// prefix as target. We need to rule out one of them to avoid to go
// to the wrong prefix.
ParsedInternalKey low_key;
uint32_t tmp;
uint32_t low_key_offset = GetFixed32Element(base_ptr, low);
Status s = decoder->NextKeyNoValue(low_key_offset, &low_key, nullptr, &tmp);
if (!s.ok()) {
return s;
}
if (GetPrefix(low_key) == prefix) {
prefix_matched = true;
*offset = low_key_offset;
} else if (low + 1 < upper_bound) {
// There is possible a next prefix, return it
prefix_matched = false;
*offset = GetFixed32Element(base_ptr, low + 1);
} else {
// target is larger than a key of the last prefix in this bucket
// but with a different prefix. Key does not exist.
*offset = file_info_.data_end_offset;
}
return Status::OK();
}
bool PlainTableReader::MatchBloom(uint32_t hash) const {
if (!enable_bloom_) {
return true;
}
if (bloom_.MayContainHash(hash)) {
PERF_COUNTER_ADD(bloom_sst_hit_count, 1);
return true;
} else {
PERF_COUNTER_ADD(bloom_sst_miss_count, 1);
return false;
}
}
Status PlainTableReader::Next(PlainTableKeyDecoder* decoder, uint32_t* offset,
ParsedInternalKey* parsed_key,
Slice* internal_key, Slice* value,
bool* seekable) const {
if (*offset == file_info_.data_end_offset) {
*offset = file_info_.data_end_offset;
return Status::OK();
}
if (*offset > file_info_.data_end_offset) {
return Status::Corruption("Offset is out of file size");
}
uint32_t bytes_read;
Status s = decoder->NextKey(*offset, parsed_key, internal_key, value,
&bytes_read, seekable);
if (!s.ok()) {
return s;
}
*offset = *offset + bytes_read;
return Status::OK();
}
void PlainTableReader::Prepare(const Slice& target) {
if (enable_bloom_) {
uint32_t prefix_hash = GetSliceHash(GetPrefix(target));
bloom_.Prefetch(prefix_hash);
}
}
Status PlainTableReader::Get(const ReadOptions& /*ro*/, const Slice& target,
GetContext* get_context,
const SliceTransform* /* prefix_extractor */,
bool /*skip_filters*/) {
// Check bloom filter first.
Slice prefix_slice;
uint32_t prefix_hash;
if (IsTotalOrderMode()) {
if (full_scan_mode_) {
status_ =
Status::InvalidArgument("Get() is not allowed in full scan mode.");
}
// Match whole user key for bloom filter check.
if (!MatchBloom(GetSliceHash(GetUserKey(target)))) {
return Status::OK();
}
// in total order mode, there is only one bucket 0, and we always use empty
// prefix.
prefix_slice = Slice();
prefix_hash = 0;
} else {
prefix_slice = GetPrefix(target);
prefix_hash = GetSliceHash(prefix_slice);
if (!MatchBloom(prefix_hash)) {
return Status::OK();
}
}
uint32_t offset;
bool prefix_match;
PlainTableKeyDecoder decoder(&file_info_, encoding_type_, user_key_len_,
prefix_extractor_);
Status s = GetOffset(&decoder, target, prefix_slice, prefix_hash,
prefix_match, &offset);
if (!s.ok()) {
return s;
}
ParsedInternalKey found_key;
ParsedInternalKey parsed_target;
if (!ParseInternalKey(target, &parsed_target)) {
return Status::Corruption(Slice());
}
Slice found_value;
while (offset < file_info_.data_end_offset) {
s = Next(&decoder, &offset, &found_key, nullptr, &found_value);
if (!s.ok()) {
return s;
}
if (!prefix_match) {
// Need to verify prefix for the first key found if it is not yet
// checked.
if (GetPrefix(found_key) != prefix_slice) {
return Status::OK();
}
prefix_match = true;
}
// TODO(ljin): since we know the key comparison result here,
// can we enable the fast path?
if (internal_comparator_.Compare(found_key, parsed_target) >= 0) {
bool dont_care __attribute__((__unused__));
if (!get_context->SaveValue(found_key, found_value, &dont_care,
dummy_cleanable_.get())) {
break;
}
}
}
return Status::OK();
}
uint64_t PlainTableReader::ApproximateOffsetOf(const Slice& /*key*/) {
return 0;
}
PlainTableIterator::PlainTableIterator(PlainTableReader* table,
bool use_prefix_seek)
: table_(table),
decoder_(&table_->file_info_, table_->encoding_type_,
table_->user_key_len_, table_->prefix_extractor_),
use_prefix_seek_(use_prefix_seek) {
next_offset_ = offset_ = table_->file_info_.data_end_offset;
}
PlainTableIterator::~PlainTableIterator() {
}
bool PlainTableIterator::Valid() const {
return offset_ < table_->file_info_.data_end_offset &&
offset_ >= table_->data_start_offset_;
}
void PlainTableIterator::SeekToFirst() {
status_ = Status::OK();
next_offset_ = table_->data_start_offset_;
if (next_offset_ >= table_->file_info_.data_end_offset) {
next_offset_ = offset_ = table_->file_info_.data_end_offset;
} else {
Next();
}
}
void PlainTableIterator::SeekToLast() {
assert(false);
status_ = Status::NotSupported("SeekToLast() is not supported in PlainTable");
next_offset_ = offset_ = table_->file_info_.data_end_offset;
}
void PlainTableIterator::Seek(const Slice& target) {
if (use_prefix_seek_ != !table_->IsTotalOrderMode()) {
// This check is done here instead of NewIterator() to permit creating an
// iterator with total_order_seek = true even if we won't be able to Seek()
// it. This is needed for compaction: it creates iterator with
// total_order_seek = true but usually never does Seek() on it,
// only SeekToFirst().
status_ =
Status::InvalidArgument(
"total_order_seek not implemented for PlainTable.");
offset_ = next_offset_ = table_->file_info_.data_end_offset;
return;
}
// If the user doesn't set prefix seek option and we are not able to do a
// total Seek(). assert failure.
if (table_->IsTotalOrderMode()) {
if (table_->full_scan_mode_) {
status_ =
Status::InvalidArgument("Seek() is not allowed in full scan mode.");
offset_ = next_offset_ = table_->file_info_.data_end_offset;
return;
} else if (table_->GetIndexSize() > 1) {
assert(false);
status_ = Status::NotSupported(
"PlainTable cannot issue non-prefix seek unless in total order "
"mode.");
offset_ = next_offset_ = table_->file_info_.data_end_offset;
return;
}
}
Slice prefix_slice = table_->GetPrefix(target);
uint32_t prefix_hash = 0;
// Bloom filter is ignored in total-order mode.
if (!table_->IsTotalOrderMode()) {
prefix_hash = GetSliceHash(prefix_slice);
if (!table_->MatchBloom(prefix_hash)) {
status_ = Status::OK();
offset_ = next_offset_ = table_->file_info_.data_end_offset;
return;
}
}
bool prefix_match;
status_ = table_->GetOffset(&decoder_, target, prefix_slice, prefix_hash,
prefix_match, &next_offset_);
if (!status_.ok()) {
offset_ = next_offset_ = table_->file_info_.data_end_offset;
return;
}
if (next_offset_ < table_->file_info_.data_end_offset) {
for (Next(); status_.ok() && Valid(); Next()) {
if (!prefix_match) {
// Need to verify the first key's prefix
if (table_->GetPrefix(key()) != prefix_slice) {
offset_ = next_offset_ = table_->file_info_.data_end_offset;
break;
}
prefix_match = true;
}
if (table_->internal_comparator_.Compare(key(), target) >= 0) {
break;
}
}
} else {
offset_ = table_->file_info_.data_end_offset;
}
}
void PlainTableIterator::SeekForPrev(const Slice& /*target*/) {
assert(false);
status_ =
Status::NotSupported("SeekForPrev() is not supported in PlainTable");
offset_ = next_offset_ = table_->file_info_.data_end_offset;
}
void PlainTableIterator::Next() {
offset_ = next_offset_;
if (offset_ < table_->file_info_.data_end_offset) {
Slice tmp_slice;
ParsedInternalKey parsed_key;
status_ =
table_->Next(&decoder_, &next_offset_, &parsed_key, &key_, &value_);
if (!status_.ok()) {
offset_ = next_offset_ = table_->file_info_.data_end_offset;
}
}
}
void PlainTableIterator::Prev() {
assert(false);
}
Slice PlainTableIterator::key() const {
assert(Valid());
return key_;
}
Slice PlainTableIterator::value() const {
assert(Valid());
return value_;
}
Status PlainTableIterator::status() const {
return status_;
}
} // namespace rocksdb
#endif // ROCKSDB_LITE
|