0af157f9bf
Summary:
1. Make filter_block.h a base class. Derive block_based_filter_block and full_filter_block. The previous one is the traditional filter block. The full_filter_block is newly added. It would generate a filter block that contain all the keys in SST file.
2. When querying a key, table would first check if full_filter is available. If not, it would go to the exact data block and check using block_based filter.
3. User could choose to use full_filter or tradional(block_based_filter). They would be stored in SST file with different meta index name. "filter.filter_policy" or "full_filter.filter_policy". Then, Table reader is able to know the fllter block type.
4. Some optimizations have been done for full_filter_block, thus it requires a different interface compared to the original one in filter_policy.h.
5. Actual implementation of filter bits coding/decoding is placed in util/bloom_impl.cc
Benchmark: base commit 1d23b5c470
Command:
db_bench --db=/dev/shm/rocksdb --num_levels=6 --key_size=20 --prefix_size=20 --keys_per_prefix=0 --value_size=100 --write_buffer_size=134217728 --max_write_buffer_number=2 --target_file_size_base=33554432 --max_bytes_for_level_base=1073741824 --verify_checksum=false --max_background_compactions=4 --use_plain_table=0 --memtablerep=prefix_hash --open_files=-1 --mmap_read=1 --mmap_write=0 --bloom_bits=10 --bloom_locality=1 --memtable_bloom_bits=500000 --compression_type=lz4 --num=393216000 --use_hash_search=1 --block_size=1024 --block_restart_interval=16 --use_existing_db=1 --threads=1 --benchmarks=readrandom —disable_auto_compactions=1
Read QPS increase for about 30% from 2230002 to 2991411.
Test Plan:
make all check
valgrind db_test
db_stress --use_block_based_filter = 0
./auto_sanity_test.sh
Reviewers: igor, yhchiang, ljin, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D20979
100 lines
2.8 KiB
C++
100 lines
2.8 KiB
C++
// Copyright (c) 2014, Facebook, Inc. All rights reserved.
|
|
// This source code is licensed under the BSD-style license found in the
|
|
// LICENSE file in the root directory of this source tree. An additional grant
|
|
// of patent rights can be found in the PATENTS file in the same directory.
|
|
|
|
#include "table/full_filter_block.h"
|
|
|
|
#include "rocksdb/filter_policy.h"
|
|
#include "port/port.h"
|
|
#include "util/coding.h"
|
|
|
|
namespace rocksdb {
|
|
|
|
FullFilterBlockBuilder::FullFilterBlockBuilder(
|
|
const SliceTransform* prefix_extractor,
|
|
const BlockBasedTableOptions& table_opt,
|
|
FilterBitsBuilder* filter_bits_builder)
|
|
: prefix_extractor_(prefix_extractor),
|
|
whole_key_filtering_(table_opt.whole_key_filtering),
|
|
num_added_(0) {
|
|
assert(filter_bits_builder != nullptr);
|
|
filter_bits_builder_.reset(filter_bits_builder);
|
|
}
|
|
|
|
void FullFilterBlockBuilder::Add(const Slice& key) {
|
|
if (whole_key_filtering_) {
|
|
AddKey(key);
|
|
}
|
|
if (prefix_extractor_ && prefix_extractor_->InDomain(key)) {
|
|
AddPrefix(key);
|
|
}
|
|
}
|
|
|
|
// Add key to filter if needed
|
|
inline void FullFilterBlockBuilder::AddKey(const Slice& key) {
|
|
filter_bits_builder_->AddKey(key);
|
|
num_added_++;
|
|
}
|
|
|
|
// Add prefix to filter if needed
|
|
inline void FullFilterBlockBuilder::AddPrefix(const Slice& key) {
|
|
Slice prefix = prefix_extractor_->Transform(key);
|
|
filter_bits_builder_->AddKey(prefix);
|
|
num_added_++;
|
|
}
|
|
|
|
Slice FullFilterBlockBuilder::Finish() {
|
|
if (num_added_ != 0) {
|
|
num_added_ = 0;
|
|
return filter_bits_builder_->Finish(&filter_data);
|
|
}
|
|
return Slice();
|
|
}
|
|
|
|
FullFilterBlockReader::FullFilterBlockReader(
|
|
const SliceTransform* prefix_extractor,
|
|
const BlockBasedTableOptions& table_opt,
|
|
const Slice& contents,
|
|
FilterBitsReader* filter_bits_reader, bool delete_contents_after_use)
|
|
: prefix_extractor_(prefix_extractor),
|
|
whole_key_filtering_(table_opt.whole_key_filtering),
|
|
contents_(contents) {
|
|
assert(filter_bits_reader != nullptr);
|
|
filter_bits_reader_.reset(filter_bits_reader);
|
|
|
|
if (delete_contents_after_use) {
|
|
filter_data.reset(contents.data());
|
|
}
|
|
}
|
|
|
|
bool FullFilterBlockReader::KeyMayMatch(const Slice& key,
|
|
uint64_t block_offset) {
|
|
assert(block_offset == kNotValid);
|
|
if (!whole_key_filtering_) {
|
|
return true;
|
|
}
|
|
return MayMatch(key);
|
|
}
|
|
|
|
bool FullFilterBlockReader::PrefixMayMatch(const Slice& prefix,
|
|
uint64_t block_offset) {
|
|
assert(block_offset == kNotValid);
|
|
if (!prefix_extractor_) {
|
|
return true;
|
|
}
|
|
return MayMatch(prefix);
|
|
}
|
|
|
|
bool FullFilterBlockReader::MayMatch(const Slice& entry) {
|
|
if (contents_.size() != 0) {
|
|
return filter_bits_reader_->MayMatch(entry);
|
|
}
|
|
return true; // remain the same with block_based filter
|
|
}
|
|
|
|
size_t FullFilterBlockReader::ApproximateMemoryUsage() const {
|
|
return contents_.size();
|
|
}
|
|
} // namespace rocksdb
|