Key and value record structure in the file hash database. More...

#include <tkrzw_dbm_hash_impl.h>

Public Types
enum	OperationType : int32_t { OP_VOID = 0 , OP_SET = 1 , OP_REMOVE = 2 , OP_ADD = 3 }
	Enumeration for operation types. More...

Public Member Functions
	HashRecord (File *file, int32_t crc_width, int32_t offset_width, int32_t align_pow)
	Constructor. More...

	~HashRecord ()
	Destructor. More...

OperationType	GetOperationType () const
	Gets the operation type of the record. More...

std::string_view	GetKey () const
	Gets the key data. More...

std::string_view	GetValue () const
	Gets the value data. More...

int64_t	GetChildOffset () const
	Gets the offset of the child record. More...

int32_t	GetWholeSize () const
	Gets the whole size of the record. More...

Status	ReadMetadataKey (int64_t offset, int32_t min_read_size)
	Reads the metadata and the key. More...

Status	ReadBody ()
	Reads the body data and fill all the properties. More...

void	SetBodyWithNullValue ()
	Sets the body data with the value filled with null codes. More...

Status	CheckCRC ()
	Check the CRC. More...

void	SetData (OperationType type, int32_t ideal_whole_size, const char key_ptr, int32_t key_size, const char value_ptr, int32_t value_size, int64_t child_offset)
	Sets the actual data of the record. More...

Status	Write (int64_t offset, int64_t *new_offset) const
	Writes the record in the file. More...

Status	WriteChildOffset (int64_t offset, int64_t child_offset)
	Writes the child offset of the record in the file. More...

Status	FindNextOffset (int64_t offset, int32_t min_read_size, int64_t *next_offset)
	Finds the next record offset by heuristics. More...

Static Public Member Functions
static Status	ReplayOperations (File file, DBM::RecordProcessor proc, int64_t record_base, int32_t crc_width, Compressor *compressor, int32_t offset_width, int32_t align_pow, int32_t min_read_size, bool skip_broken_records, int64_t end_offset)
	Replays operations applied on a hash database file. More...

static Status	ExtractOffsets (File in_file, File out_file, int64_t record_base, int32_t crc_width, int32_t offset_width, int32_t align_pow, bool skip_broken_records, int64_t end_offset)
	Extracts a sequence of offsets from a file. More...

Static Public Attributes
static constexpr int64_t	MAX_KEY_SIZE = 8 * (1LL << 20) - 1
	The maximum key size. More...

static constexpr int64_t	MAX_VALUE_SIZE = 2 * (1LL << 30) - 1
	The maximum value size. More...

static constexpr int64_t	META_MIN_READ_SIZE = 48
	The minimum size to read the meta data. More...

Detailed Description

Key and value record structure in the file hash database.

Member Enumeration Documentation

◆ OperationType

enum tkrzw::HashRecord::OperationType : int32_t

Enumeration for operation types.

Enumerator
OP_VOID	Operation to do nothing.
OP_SET	Operation to set the value of an existing record.
OP_REMOVE	Operation to remove an existing record.
OP_ADD	Operation to add a new record.

Constructor & Destructor Documentation

◆ HashRecord()

tkrzw::HashRecord::HashRecord	(	File *	file,
		int32_t	crc_width,
		int32_t	offset_width,
		int32_t	align_pow
	)

Constructor.

Parameters

file	The pointer to the file object.
crc_width	The width of the CRC value.
offset_width	The width of the offset data.
align_pow	The alignment power.

◆ ~HashRecord()

tkrzw::HashRecord::~HashRecord ( )

Destructor.

Member Function Documentation

◆ GetOperationType()

OperationType tkrzw::HashRecord::GetOperationType ( ) const

Gets the operation type of the record.

Returns: The operation type of the record.

◆ GetKey()

std::string_view tkrzw::HashRecord::GetKey ( ) const

Gets the key data.

Returns: The key data.

◆ GetValue()

std::string_view tkrzw::HashRecord::GetValue ( ) const

Gets the value data.

Returns: The value data. The data might be nullptr if the body hasn't been read.

◆ GetChildOffset()

int64_t tkrzw::HashRecord::GetChildOffset ( ) const

Gets the offset of the child record.

Returns: The offset of the child record.

◆ GetWholeSize()

int32_t tkrzw::HashRecord::GetWholeSize ( ) const

Gets the whole size of the record.

Returns: The whole size of the record. It might be zero if the body hasn't been read.

◆ ReadMetadataKey()

Status tkrzw::HashRecord::ReadMetadataKey	(	int64_t	offset,
		int32_t	min_read_size
	)

Reads the metadata and the key.

Parameters

offset	The offset of the record.
min_read_size	The minimum reading size.

Returns: The result status.

If successful, the key data is always read. However the value data and the whole size is not always read. To read them, call ReadBody.

◆ ReadBody()

Status tkrzw::HashRecord::ReadBody ( )

Reads the body data and fill all the properties.

Returns: The result status.

◆ SetBodyWithNullValue()

void tkrzw::HashRecord::SetBodyWithNullValue ( )

Sets the body data with the value filled with null codes.

◆ CheckCRC()

Status tkrzw::HashRecord::CheckCRC ( )

Check the CRC.

Returns: The result status.

◆ SetData()

void tkrzw::HashRecord::SetData	(	OperationType	type,
		int32_t	ideal_whole_size,
		const char *	key_ptr,
		int32_t	key_size,
		const char *	value_ptr,
		int32_t	value_size,
		int64_t	child_offset
	)

Sets the actual data of the record.

Parameters

type	An operation type.
ideal_whole_size	The ideal size of the storage space.
key_ptr	The pointer to a key buffer.
key_size	The size of the key buffer.
value_ptr	The pointer to a value buffer.
value_size	The size of the value buffer.
child_offset	The offset of the child record or zero for nothing.

◆ Write()

Status tkrzw::HashRecord::Write	(	int64_t	offset,
		int64_t *	new_offset
	)		const

Writes the record in the file.

Parameters

offset	The offset of the record. If it is negative, the data is appended at the end of the file.
new_offset	The pointer to an integer to store the offset of the appended data. If can be nullptr if the given offset is not negative.

Returns: The result status.

◆ WriteChildOffset()

Status tkrzw::HashRecord::WriteChildOffset	(	int64_t	offset,
		int64_t	child_offset
	)

Writes the child offset of the record in the file.

Parameters

offset	The offset of the record to update.
child_offset	The offset of the child record.

Returns: The result status.

◆ FindNextOffset()

Status tkrzw::HashRecord::FindNextOffset	(	int64_t	offset,
		int32_t	min_read_size,
		int64_t *	next_offset
	)

Finds the next record offset by heuristics.

Parameters

offset	The current offset.
min_read_size	The minimum reading size.
next_offset	The pointer to an integer to store the next offset.

Returns: The result status.

◆ ReplayOperations()

static Status tkrzw::HashRecord::ReplayOperations	(	File *	file,
		DBM::RecordProcessor *	proc,
		int64_t	record_base,
		int32_t	crc_width,
		Compressor *	compressor,
		int32_t	offset_width,
		int32_t	align_pow,
		int32_t	min_read_size,
		bool	skip_broken_records,
		int64_t	end_offset
	)

static

Replays operations applied on a hash database file.

Parameters

file	A file object having opened the database file.
proc	The pointer to the processor object.
record_base	The record base offset.
crc_width	The width of the CRC value.
compressor	The compressor object.
offset_width	The offset width.
align_pow	The alignment power.
min_read_size	The minimum reading size.
skip_broken_records	If true, the operation continues even if there are broken records which can be skipped.
end_offset	The exclusive end offset of records to read. Negative means unlimited.

Returns: The result status.

For each setting operation, ProcessFull of the processer is called. For each removing operation, ProcessEmpty of the processor is called. If they return a value other than NOOP, the iteration is cancelled.

◆ ExtractOffsets()

static Status tkrzw::HashRecord::ExtractOffsets	(	File *	in_file,
		File *	out_file,
		int64_t	record_base,
		int32_t	crc_width,
		int32_t	offset_width,
		int32_t	align_pow,
		bool	skip_broken_records,
		int64_t	end_offset
	)

static

Extracts a sequence of offsets from a file.

Parameters

in_file	A file object having opened the input database file.
out_file	A file object having opened as a writer to store the output.
record_base	The record base offset.
crc_width	The width of the CRC value.
offset_width	The offset width.
align_pow	The alignment power.
skip_broken_records	If true, the operation continues even if there are broken records which can be skipped.
end_offset	The exclusive end offset of records to read. Negative means unlimited.

Returns: The result status.

Member Data Documentation

◆ MAX_KEY_SIZE

constexpr int64_t tkrzw::HashRecord::MAX_KEY_SIZE = 8 * (1LL << 20) - 1

staticconstexpr

The maximum key size.

◆ MAX_VALUE_SIZE

constexpr int64_t tkrzw::HashRecord::MAX_VALUE_SIZE = 2 * (1LL << 30) - 1

staticconstexpr

The maximum value size.

◆ META_MIN_READ_SIZE

constexpr int64_t tkrzw::HashRecord::META_MIN_READ_SIZE = 48

staticconstexpr

The minimum size to read the meta data.

Public Types

Public Member Functions

Static Public Member Functions

Static Public Attributes

Detailed Description

Member Enumeration Documentation

◆ OperationType

Constructor & Destructor Documentation

◆ HashRecord()

◆ ~HashRecord()

Member Function Documentation

◆ GetOperationType()

◆ GetKey()

◆ GetValue()

◆ GetChildOffset()

◆ GetWholeSize()

◆ ReadMetadataKey()

◆ ReadBody()

◆ SetBodyWithNullValue()

◆ CheckCRC()

◆ SetData()

◆ Write()

◆ WriteChildOffset()

◆ FindNextOffset()

◆ ReplayOperations()

◆ ExtractOffsets()

Member Data Documentation

◆ MAX_KEY_SIZE

◆ MAX_VALUE_SIZE

◆ META_MIN_READ_SIZE