You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
rocksdb/table/block_based/block_builder.h

129 lines
5.6 KiB

// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
//
// 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.
#pragma once
#include <stdint.h>
#include <vector>
#include "rocksdb/slice.h"
#include "rocksdb/table.h"
#include "table/block_based/data_block_hash_index.h"
namespace ROCKSDB_NAMESPACE {
class BlockBuilder {
public:
BlockBuilder(const BlockBuilder&) = delete;
void operator=(const BlockBuilder&) = delete;
Introduce ReadOptions::pin_data (support zero copy for keys) Summary: This patch update the Iterator API to introduce new functions that allow users to keep the Slices returned by key() valid as long as the Iterator is not deleted ReadOptions::pin_data : If true keep loaded blocks in memory as long as the iterator is not deleted Iterator::IsKeyPinned() : If true, this mean that the Slice returned by key() is valid as long as the iterator is not deleted Also add a new option BlockBasedTableOptions::use_delta_encoding to allow users to disable delta_encoding if needed. Benchmark results (using https://phabricator.fb.com/P20083553) ``` // $ du -h /home/tec/local/normal.4K.Snappy/db10077 // 6.1G /home/tec/local/normal.4K.Snappy/db10077 // $ du -h /home/tec/local/zero.8K.LZ4/db10077 // 6.4G /home/tec/local/zero.8K.LZ4/db10077 // Benchmarks for shard db10077 // _build/opt/rocks/benchmark/rocks_copy_benchmark \ // --normal_db_path="/home/tec/local/normal.4K.Snappy/db10077" \ // --zero_db_path="/home/tec/local/zero.8K.LZ4/db10077" // First run // ============================================================================ // rocks/benchmark/RocksCopyBenchmark.cpp relative time/iter iters/s // ============================================================================ // BM_StringCopy 1.73s 576.97m // BM_StringPiece 103.74% 1.67s 598.55m // ============================================================================ // Match rate : 1000000 / 1000000 // Second run // ============================================================================ // rocks/benchmark/RocksCopyBenchmark.cpp relative time/iter iters/s // ============================================================================ // BM_StringCopy 611.99ms 1.63 // BM_StringPiece 203.76% 300.35ms 3.33 // ============================================================================ // Match rate : 1000000 / 1000000 ``` Test Plan: Unit tests Reviewers: sdong, igor, anthony, yhchiang, rven Reviewed By: rven Subscribers: dhruba, lovro, adsharma Differential Revision: https://reviews.facebook.net/D48999
9 years ago
explicit BlockBuilder(int block_restart_interval,
bool use_delta_encoding = true,
bool use_value_delta_encoding = false,
BlockBasedTableOptions::DataBlockIndexType index_type =
BlockBasedTableOptions::kDataBlockBinarySearch,
double data_block_hash_table_util_ratio = 0.75,
size_t ts_sz = 0,
bool persist_user_defined_timestamps = true,
bool is_user_key = false);
// Reset the contents as if the BlockBuilder was just constructed.
void Reset();
// Swap the contents in BlockBuilder with buffer, then reset the BlockBuilder.
void SwapAndReset(std::string& buffer);
// REQUIRES: Finish() has not been called since the last call to Reset().
// REQUIRES: Unless a range tombstone block, key is larger than any previously
// added key
// DO NOT mix with AddWithLastKey() between Resets. For efficiency, use
// AddWithLastKey() in contexts where previous added key is already known
// and delta encoding might be used.
void Add(const Slice& key, const Slice& value,
const Slice* const delta_value = nullptr);
// A faster version of Add() if the previous key is already known for all
// Add()s.
// REQUIRES: Finish() has not been called since the last call to Reset().
// REQUIRES: Unless a range tombstone block, key is larger than any previously
// added key
// REQUIRES: if AddWithLastKey has been called since last Reset(), last_key
// is the key from most recent AddWithLastKey. (For convenience, last_key
// is ignored on first call after creation or Reset().)
// DO NOT mix with Add() between Resets.
void AddWithLastKey(const Slice& key, const Slice& value,
const Slice& last_key,
const Slice* const delta_value = nullptr);
// Finish building the block and return a slice that refers to the
// block contents. The returned slice will remain valid for the
// lifetime of this builder or until Reset() is called.
Slice Finish();
// Returns an estimate of the current (uncompressed) size of the block
// we are building.
inline size_t CurrentSizeEstimate() const {
return estimate_ + (data_block_hash_index_builder_.Valid()
? data_block_hash_index_builder_.EstimateSize()
: 0);
}
// Returns an estimated block size after appending key and value.
size_t EstimateSizeAfterKV(const Slice& key, const Slice& value) const;
// Return true iff no entries have been added since the last Reset()
bool empty() const { return buffer_.empty(); }
private:
inline void AddWithLastKeyImpl(const Slice& key, const Slice& value,
const Slice& last_key,
const Slice* const delta_value,
size_t buffer_size);
inline const Slice MaybeStripTimestampFromKey(std::string* key_buf,
const Slice& key);
const int block_restart_interval_;
// TODO(myabandeh): put it into a separate IndexBlockBuilder
const bool use_delta_encoding_;
// Refer to BlockIter::DecodeCurrentValue for format of delta encoded values
const bool use_value_delta_encoding_;
Logically strip timestamp during flush (#11557) Summary: Logically strip the user-defined timestamp when L0 files are created during flush when `AdvancedColumnFamilyOptions.persist_user_defined_timestamps` is false. Logically stripping timestamp here means replacing the original user-defined timestamp with a mininum timestamp, which for now is hard coded to be all zeros bytes. While working on this, I caught a missing piece on the `BlockBuilder` level for this feature. The current quick path `std::min(buffer_size, last_key_size)` needs a bit tweaking to work for this feature. When user-defined timestamp is stripped during block building, on writing first entry or right after resetting, `buffer` is empty and `buffer_size` is zero as usual. However, in follow-up writes, depending on the size of the stripped user-defined timestamp, and the size of the value, what's in `buffer` can sometimes be smaller than `last_key_size`, leading `std::min(buffer_size, last_key_size)` to truncate the `last_key`. Previous test doesn't caught the bug because in those tests, the size of the stripped user-defined timestamps bytes is smaller than the length of the value. In order to avoid the conditional operation, this PR changed the original trivial `std::min` operation into an arithmetic operation. Since this is a change in a hot and performance critical path, I did the following benchmark to check no observable regression is introduced. ```TEST_TMPDIR=/dev/shm/rocksdb1 ./db_bench -benchmarks=fillseq -memtablerep=vector -allow_concurrent_memtable_write=false -num=50000000``` Compiled with DEBUG_LEVEL=0 Test vs. control runs simulaneous for better accuracy, units = ops/sec PR vs base: Round 1: 350652 vs 349055 Round 2: 365733 vs 364308 Round 3: 355681 vs 354475 Pull Request resolved: https://github.com/facebook/rocksdb/pull/11557 Test Plan: New timestamp specific test added or existing tests augmented, both are parameterized with `UserDefinedTimestampTestMode`: `UserDefinedTimestampTestMode::kNormal` -> UDT feature enabled, write / read with min timestamp `UserDefinedTimestampTestMode::kStripUserDefinedTimestamps` -> UDT feature enabled, write / read with min timestamp, set Options.persist_user_defined_timestamps to false. ``` make all check ./db_wal_test --gtest_filter="*WithTimestamp*" ./flush_job_test --gtest_filter="*WithTimestamp*" ./repair_test --gtest_filter="*WithTimestamp*" ./block_based_table_reader_test ``` Reviewed By: pdillinger Differential Revision: D47027664 Pulled By: jowlyzhang fbshipit-source-id: e729193b6334dfc63aaa736d684d907a022571f5
1 year ago
// Size in bytes for the user-defined timestamp to strip in a user key.
// This is non-zero if there is user-defined timestamp in the user key and it
// should not be persisted.
const size_t strip_ts_sz_;
// Whether the keys provided to build this block are user keys. If not,
// the keys are internal keys. This will affect how timestamp stripping is
// done for the key if `persisted_user_defined_timestamps_` is false and
// `ts_sz_` is non-zero.
// The timestamp stripping only applies to the keys added to the block. If the
// value contains user defined timestamp that needed to be stripped too, such
// as the `first_internal_key` in an `IndexValue` for an index block, the
// value part for a range deletion entry, their timestamp should be stripped
// before calling `BlockBuilder::Add`.
// Timestamp stripping only applies to data block and index blocks including
// index block for data blocks, index block for partitioned filter blocks,
// index block for partitioned index blocks. In summary, this only applies to
// block whose key are real user keys or internal keys created from user keys.
const bool is_user_key_;
std::string buffer_; // Destination buffer
std::vector<uint32_t> restarts_; // Restart points
size_t estimate_;
int counter_; // Number of entries emitted since restart
bool finished_; // Has Finish() been called?
std::string last_key_;
DataBlockHashIndexBuilder data_block_hash_index_builder_;
#ifndef NDEBUG
bool add_with_last_key_called_ = false;
#endif
};
} // namespace ROCKSDB_NAMESPACE