2016-04-01 21:42:24 +08:00
|
|
|
/*-------------------------------------------------------------------------
|
|
|
|
*
|
|
|
|
* bloom.h
|
|
|
|
* Header for bloom index.
|
|
|
|
*
|
2022-01-08 08:04:57 +08:00
|
|
|
* Copyright (c) 2016-2022, PostgreSQL Global Development Group
|
2016-04-01 21:42:24 +08:00
|
|
|
*
|
|
|
|
* IDENTIFICATION
|
|
|
|
* contrib/bloom/bloom.h
|
|
|
|
*
|
|
|
|
*-------------------------------------------------------------------------
|
|
|
|
*/
|
|
|
|
#ifndef _BLOOM_H_
|
|
|
|
#define _BLOOM_H_
|
|
|
|
|
2019-12-27 07:09:00 +08:00
|
|
|
#include "access/amapi.h"
|
2016-04-01 21:42:24 +08:00
|
|
|
#include "access/generic_xlog.h"
|
|
|
|
#include "access/itup.h"
|
|
|
|
#include "access/xlog.h"
|
|
|
|
#include "fmgr.h"
|
2019-11-25 10:38:57 +08:00
|
|
|
#include "nodes/pathnodes.h"
|
2016-04-01 21:42:24 +08:00
|
|
|
|
|
|
|
/* Support procedures numbers */
|
|
|
|
#define BLOOM_HASH_PROC 1
|
Implement operator class parameters
PostgreSQL provides set of template index access methods, where opclasses have
much freedom in the semantics of indexing. These index AMs are GiST, GIN,
SP-GiST and BRIN. There opclasses define representation of keys, operations on
them and supported search strategies. So, it's natural that opclasses may be
faced some tradeoffs, which require user-side decision. This commit implements
opclass parameters allowing users to set some values, which tell opclass how to
index the particular dataset.
This commit doesn't introduce new storage in system catalog. Instead it uses
pg_attribute.attoptions, which is used for table column storage options but
unused for index attributes.
In order to evade changing signature of each opclass support function, we
implement unified way to pass options to opclass support functions. Options
are set to fn_expr as the constant bytea expression. It's possible due to the
fact that opclass support functions are executed outside of expressions, so
fn_expr is unused for them.
This commit comes with some examples of opclass options usage. We parametrize
signature length in GiST. That applies to multiple opclasses: tsvector_ops,
gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and
gist_hstore_ops. Also we parametrize maximum number of integer ranges for
gist__int_ops. However, the main future usage of this feature is expected
to be json, where users would be able to specify which way to index particular
json parts.
Catversion is bumped.
Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru
Author: Nikita Glukhov, revised by me
Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-31 00:17:11 +08:00
|
|
|
#define BLOOM_OPTIONS_PROC 2
|
|
|
|
#define BLOOM_NPROC 2
|
2016-04-01 21:42:24 +08:00
|
|
|
|
|
|
|
/* Scan strategies */
|
|
|
|
#define BLOOM_EQUAL_STRATEGY 1
|
|
|
|
#define BLOOM_NSTRATEGIES 1
|
|
|
|
|
|
|
|
/* Opaque for bloom pages */
|
|
|
|
typedef struct BloomPageOpaqueData
|
|
|
|
{
|
2016-04-12 23:03:01 +08:00
|
|
|
OffsetNumber maxoff; /* number of index tuples on page */
|
|
|
|
uint16 flags; /* see bit definitions below */
|
|
|
|
uint16 unused; /* placeholder to force maxaligning of size of
|
|
|
|
* BloomPageOpaqueData and to place
|
|
|
|
* bloom_page_id exactly at the end of page */
|
|
|
|
uint16 bloom_page_id; /* for identification of BLOOM indexes */
|
2016-04-01 21:42:24 +08:00
|
|
|
} BloomPageOpaqueData;
|
|
|
|
|
|
|
|
typedef BloomPageOpaqueData *BloomPageOpaque;
|
|
|
|
|
|
|
|
/* Bloom page flags */
|
|
|
|
#define BLOOM_META (1<<0)
|
|
|
|
#define BLOOM_DELETED (2<<0)
|
|
|
|
|
2016-04-12 23:03:01 +08:00
|
|
|
/*
|
|
|
|
* The page ID is for the convenience of pg_filedump and similar utilities,
|
|
|
|
* which otherwise would have a hard time telling pages of different index
|
|
|
|
* types apart. It should be the last 2 bytes on the page. This is more or
|
|
|
|
* less "free" due to alignment considerations.
|
|
|
|
*
|
|
|
|
* See comments above GinPageOpaqueData.
|
|
|
|
*/
|
|
|
|
#define BLOOM_PAGE_ID 0xFF83
|
|
|
|
|
2016-04-01 21:42:24 +08:00
|
|
|
/* Macros for accessing bloom page structures */
|
|
|
|
#define BloomPageGetOpaque(page) ((BloomPageOpaque) PageGetSpecialPointer(page))
|
|
|
|
#define BloomPageGetMaxOffset(page) (BloomPageGetOpaque(page)->maxoff)
|
2016-04-02 01:09:13 +08:00
|
|
|
#define BloomPageIsMeta(page) \
|
|
|
|
((BloomPageGetOpaque(page)->flags & BLOOM_META) != 0)
|
|
|
|
#define BloomPageIsDeleted(page) \
|
|
|
|
((BloomPageGetOpaque(page)->flags & BLOOM_DELETED) != 0)
|
|
|
|
#define BloomPageSetDeleted(page) \
|
|
|
|
(BloomPageGetOpaque(page)->flags |= BLOOM_DELETED)
|
|
|
|
#define BloomPageSetNonDeleted(page) \
|
|
|
|
(BloomPageGetOpaque(page)->flags &= ~BLOOM_DELETED)
|
2016-04-01 21:42:24 +08:00
|
|
|
#define BloomPageGetData(page) ((BloomTuple *)PageGetContents(page))
|
|
|
|
#define BloomPageGetTuple(state, page, offset) \
|
|
|
|
((BloomTuple *)(PageGetContents(page) \
|
|
|
|
+ (state)->sizeOfBloomTuple * ((offset) - 1)))
|
|
|
|
#define BloomPageGetNextTuple(state, tuple) \
|
|
|
|
((BloomTuple *)((Pointer)(tuple) + (state)->sizeOfBloomTuple))
|
|
|
|
|
|
|
|
/* Preserved page numbers */
|
|
|
|
#define BLOOM_METAPAGE_BLKNO (0)
|
|
|
|
#define BLOOM_HEAD_BLKNO (1) /* first data page */
|
|
|
|
|
2016-04-02 18:47:04 +08:00
|
|
|
/*
|
2016-06-03 22:52:36 +08:00
|
|
|
* We store Bloom signatures as arrays of uint16 words.
|
2016-04-02 18:47:04 +08:00
|
|
|
*/
|
2016-06-03 22:52:36 +08:00
|
|
|
typedef uint16 BloomSignatureWord;
|
|
|
|
|
|
|
|
#define SIGNWORDBITS ((int) (BITS_PER_BYTE * sizeof(BloomSignatureWord)))
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Default and maximum Bloom signature length in bits.
|
|
|
|
*/
|
|
|
|
#define DEFAULT_BLOOM_LENGTH (5 * SIGNWORDBITS)
|
|
|
|
#define MAX_BLOOM_LENGTH (256 * SIGNWORDBITS)
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Default and maximum signature bits generated per index key.
|
|
|
|
*/
|
|
|
|
#define DEFAULT_BLOOM_BITS 2
|
|
|
|
#define MAX_BLOOM_BITS (MAX_BLOOM_LENGTH - 1)
|
2016-04-02 18:47:04 +08:00
|
|
|
|
2016-04-01 21:42:24 +08:00
|
|
|
/* Bloom index options */
|
|
|
|
typedef struct BloomOptions
|
|
|
|
{
|
|
|
|
int32 vl_len_; /* varlena header (do not touch directly!) */
|
2016-06-03 22:52:36 +08:00
|
|
|
int bloomLength; /* length of signature in words (not bits!) */
|
|
|
|
int bitSize[INDEX_MAX_KEYS]; /* # of bits generated for each
|
|
|
|
* index key */
|
2016-04-01 21:42:24 +08:00
|
|
|
} BloomOptions;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* FreeBlockNumberArray - array of block numbers sized so that metadata fill
|
|
|
|
* all space in metapage.
|
|
|
|
*/
|
|
|
|
typedef BlockNumber FreeBlockNumberArray[
|
|
|
|
MAXALIGN_DOWN(
|
|
|
|
BLCKSZ - SizeOfPageHeaderData - MAXALIGN(sizeof(BloomPageOpaqueData))
|
|
|
|
- MAXALIGN(sizeof(uint16) * 2 + sizeof(uint32) + sizeof(BloomOptions))
|
|
|
|
) / sizeof(BlockNumber)
|
|
|
|
];
|
|
|
|
|
|
|
|
/* Metadata of bloom index */
|
|
|
|
typedef struct BloomMetaPageData
|
|
|
|
{
|
|
|
|
uint32 magickNumber;
|
|
|
|
uint16 nStart;
|
|
|
|
uint16 nEnd;
|
|
|
|
BloomOptions opts;
|
|
|
|
FreeBlockNumberArray notFullPage;
|
|
|
|
} BloomMetaPageData;
|
|
|
|
|
|
|
|
/* Magic number to distinguish bloom pages among anothers */
|
|
|
|
#define BLOOM_MAGICK_NUMBER (0xDBAC0DED)
|
|
|
|
|
|
|
|
/* Number of blocks numbers fit in BloomMetaPageData */
|
|
|
|
#define BloomMetaBlockN (sizeof(FreeBlockNumberArray) / sizeof(BlockNumber))
|
|
|
|
|
|
|
|
#define BloomPageGetMeta(page) ((BloomMetaPageData *) PageGetContents(page))
|
|
|
|
|
|
|
|
typedef struct BloomState
|
|
|
|
{
|
|
|
|
FmgrInfo hashFn[INDEX_MAX_KEYS];
|
2019-03-22 19:09:32 +08:00
|
|
|
Oid collations[INDEX_MAX_KEYS];
|
2016-04-04 03:16:07 +08:00
|
|
|
BloomOptions opts; /* copy of options on index's metapage */
|
2016-04-01 21:42:24 +08:00
|
|
|
int32 nColumns;
|
|
|
|
|
|
|
|
/*
|
2016-04-04 03:16:07 +08:00
|
|
|
* sizeOfBloomTuple is index-specific, and it depends on reloptions, so
|
2016-04-01 21:42:24 +08:00
|
|
|
* precompute it
|
|
|
|
*/
|
2016-04-04 02:17:20 +08:00
|
|
|
Size sizeOfBloomTuple;
|
2016-04-01 21:42:24 +08:00
|
|
|
} BloomState;
|
|
|
|
|
|
|
|
#define BloomPageGetFreeSpace(state, page) \
|
|
|
|
(BLCKSZ - MAXALIGN(SizeOfPageHeaderData) \
|
|
|
|
- BloomPageGetMaxOffset(page) * (state)->sizeOfBloomTuple \
|
|
|
|
- MAXALIGN(sizeof(BloomPageOpaqueData)))
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Tuples are very different from all other relations
|
|
|
|
*/
|
|
|
|
typedef struct BloomTuple
|
|
|
|
{
|
|
|
|
ItemPointerData heapPtr;
|
2016-06-03 22:52:36 +08:00
|
|
|
BloomSignatureWord sign[FLEXIBLE_ARRAY_MEMBER];
|
2016-04-01 21:42:24 +08:00
|
|
|
} BloomTuple;
|
|
|
|
|
|
|
|
#define BLOOMTUPLEHDRSZ offsetof(BloomTuple, sign)
|
|
|
|
|
|
|
|
/* Opaque data structure for bloom index scan */
|
|
|
|
typedef struct BloomScanOpaqueData
|
|
|
|
{
|
2016-06-03 22:52:36 +08:00
|
|
|
BloomSignatureWord *sign; /* Scan signature */
|
2016-04-01 21:42:24 +08:00
|
|
|
BloomState state;
|
|
|
|
} BloomScanOpaqueData;
|
|
|
|
|
|
|
|
typedef BloomScanOpaqueData *BloomScanOpaque;
|
|
|
|
|
|
|
|
/* blutils.c */
|
|
|
|
extern void _PG_init(void);
|
|
|
|
extern void initBloomState(BloomState *state, Relation index);
|
2016-05-25 09:04:23 +08:00
|
|
|
extern void BloomFillMetapage(Relation index, Page metaPage);
|
2016-04-01 21:42:24 +08:00
|
|
|
extern void BloomInitMetapage(Relation index);
|
|
|
|
extern void BloomInitPage(Page page, uint16 flags);
|
|
|
|
extern Buffer BloomNewBuffer(Relation index);
|
2016-06-03 22:52:36 +08:00
|
|
|
extern void signValue(BloomState *state, BloomSignatureWord *sign, Datum value, int attno);
|
2016-04-01 21:42:24 +08:00
|
|
|
extern BloomTuple *BloomFormTuple(BloomState *state, ItemPointer iptr, Datum *values, bool *isnull);
|
|
|
|
extern bool BloomPageAddItem(BloomState *state, Page page, BloomTuple *tuple);
|
|
|
|
|
|
|
|
/* blvalidate.c */
|
|
|
|
extern bool blvalidate(Oid opclassoid);
|
|
|
|
|
|
|
|
/* index access method interface functions */
|
|
|
|
extern bool blinsert(Relation index, Datum *values, bool *isnull,
|
|
|
|
ItemPointer ht_ctid, Relation heapRel,
|
Allow index AMs to cache data across aminsert calls within a SQL command.
It's always been possible for index AMs to cache data across successive
amgettuple calls within a single SQL command: the IndexScanDesc.opaque
field is meant for precisely that. However, no comparable facility
exists for amortizing setup work across successive aminsert calls.
This patch adds such a feature and teaches GIN, GIST, and BRIN to use it
to amortize catalog lookups they'd previously been doing on every call.
(The other standard index AMs keep everything they need in the relcache,
so there's little to improve there.)
For GIN, the overall improvement in a statement that inserts many rows
can be as much as 10%, though it seems a bit less for the other two.
In addition, this makes a really significant difference in runtime
for CLOBBER_CACHE_ALWAYS tests, since in those builds the repeated
catalog lookups are vastly more expensive.
The reason this has been hard up to now is that the aminsert function is
not passed any useful place to cache per-statement data. What I chose to
do is to add suitable fields to struct IndexInfo and pass that to aminsert.
That's not widening the index AM API very much because IndexInfo is already
within the ken of ambuild; in fact, by passing the same info to aminsert
as to ambuild, this is really removing an inconsistency in the AM API.
Discussion: https://postgr.es/m/27568.1486508680@sss.pgh.pa.us
2017-02-10 00:52:12 +08:00
|
|
|
IndexUniqueCheck checkUnique,
|
2021-01-14 00:11:00 +08:00
|
|
|
bool indexUnchanged,
|
Allow index AMs to cache data across aminsert calls within a SQL command.
It's always been possible for index AMs to cache data across successive
amgettuple calls within a single SQL command: the IndexScanDesc.opaque
field is meant for precisely that. However, no comparable facility
exists for amortizing setup work across successive aminsert calls.
This patch adds such a feature and teaches GIN, GIST, and BRIN to use it
to amortize catalog lookups they'd previously been doing on every call.
(The other standard index AMs keep everything they need in the relcache,
so there's little to improve there.)
For GIN, the overall improvement in a statement that inserts many rows
can be as much as 10%, though it seems a bit less for the other two.
In addition, this makes a really significant difference in runtime
for CLOBBER_CACHE_ALWAYS tests, since in those builds the repeated
catalog lookups are vastly more expensive.
The reason this has been hard up to now is that the aminsert function is
not passed any useful place to cache per-statement data. What I chose to
do is to add suitable fields to struct IndexInfo and pass that to aminsert.
That's not widening the index AM API very much because IndexInfo is already
within the ken of ambuild; in fact, by passing the same info to aminsert
as to ambuild, this is really removing an inconsistency in the AM API.
Discussion: https://postgr.es/m/27568.1486508680@sss.pgh.pa.us
2017-02-10 00:52:12 +08:00
|
|
|
struct IndexInfo *indexInfo);
|
2016-04-01 21:42:24 +08:00
|
|
|
extern IndexScanDesc blbeginscan(Relation r, int nkeys, int norderbys);
|
|
|
|
extern int64 blgetbitmap(IndexScanDesc scan, TIDBitmap *tbm);
|
|
|
|
extern void blrescan(IndexScanDesc scan, ScanKey scankey, int nscankeys,
|
|
|
|
ScanKey orderbys, int norderbys);
|
|
|
|
extern void blendscan(IndexScanDesc scan);
|
|
|
|
extern IndexBuildResult *blbuild(Relation heap, Relation index,
|
|
|
|
struct IndexInfo *indexInfo);
|
|
|
|
extern void blbuildempty(Relation index);
|
|
|
|
extern IndexBulkDeleteResult *blbulkdelete(IndexVacuumInfo *info,
|
|
|
|
IndexBulkDeleteResult *stats, IndexBulkDeleteCallback callback,
|
|
|
|
void *callback_state);
|
|
|
|
extern IndexBulkDeleteResult *blvacuumcleanup(IndexVacuumInfo *info,
|
|
|
|
IndexBulkDeleteResult *stats);
|
|
|
|
extern bytea *bloptions(Datum reloptions, bool validate);
|
|
|
|
extern void blcostestimate(PlannerInfo *root, IndexPath *path,
|
|
|
|
double loop_count, Cost *indexStartupCost,
|
|
|
|
Cost *indexTotalCost, Selectivity *indexSelectivity,
|
2017-02-16 02:53:24 +08:00
|
|
|
double *indexCorrelation, double *indexPages);
|
2016-04-01 21:42:24 +08:00
|
|
|
|
|
|
|
#endif
|