Unique_on_insert – similar to above, but rejects duplicates on insert, not just on read of the final result. More...

#include <uniques.h>

Public Member Functions
	Unique_on_insert (uint size)

bool	unique_add (void *ptr)
	Add row id to the filter. More...

bool	init ()
	Initialize duplicate filter - allocate duplicate weedout tmp table. More...

void	reset (bool reinit)
	Reset filter - drop all rowid records. More...

void	cleanup ()
	Cleanup unique filter. More...

Private Attributes
uint	m_size
	Element size. More...

TABLE *	m_table {nullptr}
	Duplicate weedout tmp table. More...

Detailed Description

Unique_on_insert – similar to above, but rejects duplicates on insert, not just on read of the final result.

To achieve this values are inserted into mem tmp table which uses index to detect duplicate keys. When memory buffer is full, tmp table is dumped to a disk-based tmp table.

Constructor & Destructor Documentation

◆ Unique_on_insert()

Unique_on_insert::Unique_on_insert ( uint size )

inline

Member Function Documentation

◆ cleanup()

void Unique_on_insert::cleanup ( void )

Cleanup unique filter.

◆ init()

bool Unique_on_insert::init ( void )

Initialize duplicate filter - allocate duplicate weedout tmp table.

Returns: false initialization succeeded true an error occur

◆ reset()

void Unique_on_insert::reset ( bool reinit )

Reset filter - drop all rowid records.

Parameters

reinit Whether to restart index scan

◆ unique_add()

bool Unique_on_insert::unique_add ( void * ptr )

Add row id to the filter.

Parameters

ptr	pointer to the rowid

Returns: false rowid successfully inserted true duplicate or error

Member Data Documentation

◆ m_size

uint Unique_on_insert::m_size

private

Element size.

◆ m_table

TABLE* Unique_on_insert::m_table {nullptr}