rdelaage
/
postgresql
mirror of https://git.postgresql.org/git/postgresql.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
postgresql/src/backend/access/spgist/spgscan.c

667 lines
17 KiB
C
Raw Normal View History

Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
/*-------------------------------------------------------------------------
*
* spgscan.c
* routines for scanning SP-GiST indexes
*
*
Update copyright notices for year 2012.
2012-01-02 00:01:58 +01:00
* Portions Copyright (c) 1996-2012, PostgreSQL Global Development Group
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
* Portions Copyright (c) 1994, Regents of the University of California
*
* IDENTIFICATION
* src/backend/access/spgist/spgscan.c
*
*-------------------------------------------------------------------------
*/
#include "postgres.h"
#include "access/relscan.h"
#include "access/spgist_private.h"
#include "miscadmin.h"
#include "storage/bufmgr.h"
#include "utils/datum.h"
#include "utils/memutils.h"
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
typedef void (*storeRes_func) (SpGistScanOpaque so, ItemPointer heapPtr,
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 21:20:04 +02:00
Datum leafValue, bool isnull, bool recheck);
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
typedef struct ScanStackEntry
{
Datum reconstructedValue; /* value reconstructed from parent */
int level; /* level of items on this page */
ItemPointerData ptr; /* block and offset to scan from */
} ScanStackEntry;
/* Free a ScanStackEntry */
static void
freeScanStackEntry(SpGistScanOpaque so, ScanStackEntry *stackEntry)
{
if (!so->state.attType.attbyval &&
DatumGetPointer(stackEntry->reconstructedValue) != NULL)
pfree(DatumGetPointer(stackEntry->reconstructedValue));
pfree(stackEntry);
}
/* Free the entire stack */
static void
freeScanStack(SpGistScanOpaque so)
{
ListCell *lc;
foreach(lc, so->scanStack)
{
freeScanStackEntry(so, (ScanStackEntry *) lfirst(lc));
}
list_free(so->scanStack);
so->scanStack = NIL;
}
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
/*
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
* Initialize scanStack to search the root page, resetting
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
* any previously active scan
*/
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
static void
resetSpGistScanOpaque(SpGistScanOpaque so)
{
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
ScanStackEntry *startEntry;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
freeScanStack(so);
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
if (so->searchNulls)
{
/* Stack a work item to scan the null index entries */
startEntry = (ScanStackEntry *) palloc0(sizeof(ScanStackEntry));
ItemPointerSet(&startEntry->ptr, SPGIST_NULL_BLKNO, FirstOffsetNumber);
so->scanStack = lappend(so->scanStack, startEntry);
}
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
if (so->searchNonNulls)
{
/* Stack a work item to scan the non-null index entries */
startEntry = (ScanStackEntry *) palloc0(sizeof(ScanStackEntry));
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
ItemPointerSet(&startEntry->ptr, SPGIST_ROOT_BLKNO, FirstOffsetNumber);
so->scanStack = lappend(so->scanStack, startEntry);
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
}
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
if (so->want_itup)
{
/* Must pfree IndexTuples to avoid memory leak */
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 21:20:04 +02:00
int i;
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
for (i = 0; i < so->nPtrs; i++)
pfree(so->indexTups[i]);
}
so->iPtr = so->nPtrs = 0;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
}
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
/*
* Prepare scan keys in SpGistScanOpaque from caller-given scan keys
*
* Sets searchNulls, searchNonNulls, numberOfKeys, keyData fields of *so.
*
* The point here is to eliminate null-related considerations from what the
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 21:20:04 +02:00
* opclass consistent functions need to deal with. We assume all SPGiST-
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
* indexable operators are strict, so any null RHS value makes the scan
* condition unsatisfiable. We also pull out any IS NULL/IS NOT NULL
* conditions; their effect is reflected into searchNulls/searchNonNulls.
*/
static void
spgPrepareScanKeys(IndexScanDesc scan)
{
SpGistScanOpaque so = (SpGistScanOpaque) scan->opaque;
bool qual_ok;
bool haveIsNull;
bool haveNotNull;
int nkeys;
int i;
if (scan->numberOfKeys <= 0)
{
/* If no quals, whole-index scan is required */
so->searchNulls = true;
so->searchNonNulls = true;
so->numberOfKeys = 0;
return;
}
/* Examine the given quals */
qual_ok = true;
haveIsNull = haveNotNull = false;
nkeys = 0;
for (i = 0; i < scan->numberOfKeys; i++)
{
ScanKey skey = &scan->keyData[i];
if (skey->sk_flags & SK_SEARCHNULL)
haveIsNull = true;
else if (skey->sk_flags & SK_SEARCHNOTNULL)
haveNotNull = true;
else if (skey->sk_flags & SK_ISNULL)
{
/* ordinary qual with null argument - unsatisfiable */
qual_ok = false;
break;
}
else
{
/* ordinary qual, propagate into so->keyData */
so->keyData[nkeys++] = *skey;
/* this effectively creates a not-null requirement */
haveNotNull = true;
}
}
/* IS NULL in combination with something else is unsatisfiable */
if (haveIsNull && haveNotNull)
qual_ok = false;
/* Emit results */
if (qual_ok)
{
so->searchNulls = haveIsNull;
so->searchNonNulls = haveNotNull;
so->numberOfKeys = nkeys;
}
else
{
so->searchNulls = false;
so->searchNonNulls = false;
so->numberOfKeys = 0;
}
}
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
Datum
spgbeginscan(PG_FUNCTION_ARGS)
{
Relation rel = (Relation) PG_GETARG_POINTER(0);
int keysz = PG_GETARG_INT32(1);
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 21:20:04 +02:00
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
/* ScanKey scankey = (ScanKey) PG_GETARG_POINTER(2); */
IndexScanDesc scan;
SpGistScanOpaque so;
scan = RelationGetIndexScan(rel, keysz, 0);
so = (SpGistScanOpaque) palloc0(sizeof(SpGistScanOpaqueData));
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
if (keysz > 0)
so->keyData = (ScanKey) palloc(sizeof(ScanKeyData) * keysz);
else
so->keyData = NULL;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
initSpGistState(&so->state, scan->indexRelation);
so->tempCxt = AllocSetContextCreate(CurrentMemoryContext,
"SP-GiST search temporary context",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
/* Set up indexTupDesc and xs_itupdesc in case it's an index-only scan */
so->indexTupDesc = scan->xs_itupdesc = RelationGetDescr(rel);
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
scan->opaque = so;
PG_RETURN_POINTER(scan);
}
Datum
spgrescan(PG_FUNCTION_ARGS)
{
IndexScanDesc scan = (IndexScanDesc) PG_GETARG_POINTER(0);
SpGistScanOpaque so = (SpGistScanOpaque) scan->opaque;
ScanKey scankey = (ScanKey) PG_GETARG_POINTER(1);
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
/* copy scankeys into local storage */
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
if (scankey && scan->numberOfKeys > 0)
{
memmove(scan->keyData, scankey,
scan->numberOfKeys * sizeof(ScanKeyData));
}
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
/* preprocess scankeys, set up the representation in *so */
spgPrepareScanKeys(scan);
/* set up starting stack entries */
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
resetSpGistScanOpaque(so);
PG_RETURN_VOID();
}
Datum
spgendscan(PG_FUNCTION_ARGS)
{
IndexScanDesc scan = (IndexScanDesc) PG_GETARG_POINTER(0);
SpGistScanOpaque so = (SpGistScanOpaque) scan->opaque;
MemoryContextDelete(so->tempCxt);
PG_RETURN_VOID();
}
Datum
spgmarkpos(PG_FUNCTION_ARGS)
{
elog(ERROR, "SPGiST does not support mark/restore");
PG_RETURN_VOID();
}
Datum
spgrestrpos(PG_FUNCTION_ARGS)
{
elog(ERROR, "SPGiST does not support mark/restore");
PG_RETURN_VOID();
}
/*
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
* Test whether a leaf tuple satisfies all the scan keys
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
*
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
* *leafValue is set to the reconstructed datum, if provided
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
* *recheck is set true if any of the operators are lossy
*/
static bool
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
spgLeafTest(Relation index, SpGistScanOpaque so,
SpGistLeafTuple leafTuple, bool isnull,
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
int level, Datum reconstructedValue,
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
Datum *leafValue, bool *recheck)
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
{
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
bool result;
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
Datum leafDatum;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
spgLeafConsistentIn in;
spgLeafConsistentOut out;
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
FmgrInfo *procinfo;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
MemoryContext oldCtx;
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
if (isnull)
{
/* Should not have arrived on a nulls page unless nulls are wanted */
Assert(so->searchNulls);
*leafValue = (Datum) 0;
*recheck = false;
return true;
}
leafDatum = SGLTDATUM(leafTuple, &so->state);
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
/* use temp context for calling leaf_consistent */
oldCtx = MemoryContextSwitchTo(so->tempCxt);
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
in.scankeys = so->keyData;
in.nkeys = so->numberOfKeys;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
in.reconstructedValue = reconstructedValue;
in.level = level;
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
in.returnData = so->want_itup;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
in.leafDatum = leafDatum;
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
out.leafValue = (Datum) 0;
out.recheck = false;
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
procinfo = index_getprocinfo(index, 1, SPGIST_LEAF_CONSISTENT_PROC);
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
result = DatumGetBool(FunctionCall2Coll(procinfo,
index->rd_indcollation[0],
PointerGetDatum(&in),
PointerGetDatum(&out)));
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
*leafValue = out.leafValue;
*recheck = out.recheck;
Defend against null scankeys in spgist searches. Should've thought of that one earlier.
2011-12-18 01:08:28 +01:00
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
MemoryContextSwitchTo(oldCtx);
return result;
}
/*
* Walk the tree and report all tuples passing the scan quals to the storeRes
* subroutine.
*
* If scanWholeIndex is true, we'll do just that. If not, we'll stop at the
* next page boundary once we have reported at least one tuple.
*/
static void
spgWalk(Relation index, SpGistScanOpaque so, bool scanWholeIndex,
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
storeRes_func storeRes)
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
{
Buffer buffer = InvalidBuffer;
bool reportedSome = false;
while (scanWholeIndex || !reportedSome)
{
ScanStackEntry *stackEntry;
BlockNumber blkno;
OffsetNumber offset;
Page page;
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
bool isnull;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
/* Pull next to-do item from the list */
if (so->scanStack == NIL)
break; /* there are no more pages to scan */
stackEntry = (ScanStackEntry *) linitial(so->scanStack);
so->scanStack = list_delete_first(so->scanStack);
redirect:
/* Check for interrupts, just in case of infinite loop */
CHECK_FOR_INTERRUPTS();
blkno = ItemPointerGetBlockNumber(&stackEntry->ptr);
offset = ItemPointerGetOffsetNumber(&stackEntry->ptr);
if (buffer == InvalidBuffer)
{
buffer = ReadBuffer(index, blkno);
LockBuffer(buffer, BUFFER_LOCK_SHARE);
}
else if (blkno != BufferGetBlockNumber(buffer))
{
UnlockReleaseBuffer(buffer);
buffer = ReadBuffer(index, blkno);
LockBuffer(buffer, BUFFER_LOCK_SHARE);
}
/* else new pointer points to the same page, no work needed */
page = BufferGetPage(buffer);
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
isnull = SpGistPageStoresNulls(page) ? true : false;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
if (SpGistPageIsLeaf(page))
{
SpGistLeafTuple leafTuple;
OffsetNumber max = PageGetMaxOffsetNumber(page);
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
Datum leafValue = (Datum) 0;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
bool recheck = false;
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
if (SpGistBlockIsRoot(blkno))
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
{
/* When root is a leaf, examine all its tuples */
for (offset = FirstOffsetNumber; offset <= max; offset++)
{
leafTuple = (SpGistLeafTuple)
PageGetItem(page, PageGetItemId(page, offset));
if (leafTuple->tupstate != SPGIST_LIVE)
{
/* all tuples on root should be live */
elog(ERROR, "unexpected SPGiST tuple state: %d",
leafTuple->tupstate);
}
Assert(ItemPointerIsValid(&leafTuple->heapPtr));
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
if (spgLeafTest(index, so,
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
leafTuple, isnull,
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
stackEntry->level,
stackEntry->reconstructedValue,
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
&leafValue,
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
&recheck))
{
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
storeRes(so, &leafTuple->heapPtr,
leafValue, isnull, recheck);
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
reportedSome = true;
}
}
}
else
{
/* Normal case: just examine the chain we arrived at */
while (offset != InvalidOffsetNumber)
{
Assert(offset >= FirstOffsetNumber && offset <= max);
leafTuple = (SpGistLeafTuple)
PageGetItem(page, PageGetItemId(page, offset));
if (leafTuple->tupstate != SPGIST_LIVE)
{
if (leafTuple->tupstate == SPGIST_REDIRECT)
{
/* redirection tuple should be first in chain */
Assert(offset == ItemPointerGetOffsetNumber(&stackEntry->ptr));
/* transfer attention to redirect point */
stackEntry->ptr = ((SpGistDeadTuple) leafTuple)->pointer;
Assert(ItemPointerGetBlockNumber(&stackEntry->ptr) != SPGIST_METAPAGE_BLKNO);
goto redirect;
}
if (leafTuple->tupstate == SPGIST_DEAD)
{
/* dead tuple should be first in chain */
Assert(offset == ItemPointerGetOffsetNumber(&stackEntry->ptr));
/* No live entries on this page */
Assert(leafTuple->nextOffset == InvalidOffsetNumber);
break;
}
/* We should not arrive at a placeholder */
elog(ERROR, "unexpected SPGiST tuple state: %d",
leafTuple->tupstate);
}
Assert(ItemPointerIsValid(&leafTuple->heapPtr));
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
if (spgLeafTest(index, so,
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
leafTuple, isnull,
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
stackEntry->level,
stackEntry->reconstructedValue,
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
&leafValue,
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
&recheck))
{
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
storeRes(so, &leafTuple->heapPtr,
leafValue, isnull, recheck);
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
reportedSome = true;
}
offset = leafTuple->nextOffset;
}
}
}
else /* page is inner */
{
SpGistInnerTuple innerTuple;
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
spgInnerConsistentIn in;
spgInnerConsistentOut out;
FmgrInfo *procinfo;
SpGistNodeTuple *nodes;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
SpGistNodeTuple node;
int i;
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
MemoryContext oldCtx;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
innerTuple = (SpGistInnerTuple) PageGetItem(page,
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 21:20:04 +02:00
PageGetItemId(page, offset));
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
if (innerTuple->tupstate != SPGIST_LIVE)
{
if (innerTuple->tupstate == SPGIST_REDIRECT)
{
/* transfer attention to redirect point */
stackEntry->ptr = ((SpGistDeadTuple) innerTuple)->pointer;
Assert(ItemPointerGetBlockNumber(&stackEntry->ptr) != SPGIST_METAPAGE_BLKNO);
goto redirect;
}
elog(ERROR, "unexpected SPGiST tuple state: %d",
innerTuple->tupstate);
}
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
/* use temp context for calling inner_consistent */
oldCtx = MemoryContextSwitchTo(so->tempCxt);
in.scankeys = so->keyData;
in.nkeys = so->numberOfKeys;
in.reconstructedValue = stackEntry->reconstructedValue;
in.level = stackEntry->level;
in.returnData = so->want_itup;
in.allTheSame = innerTuple->allTheSame;
in.hasPrefix = (innerTuple->prefixSize > 0);
in.prefixDatum = SGITDATUM(innerTuple, &so->state);
in.nNodes = innerTuple->nNodes;
in.nodeLabels = spgExtractNodeLabels(&so->state, innerTuple);
/* collect node pointers */
nodes = (SpGistNodeTuple *) palloc(sizeof(SpGistNodeTuple) * in.nNodes);
SGITITERATE(innerTuple, i, node)
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
{
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
nodes[i] = node;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
}
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
memset(&out, 0, sizeof(out));
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
if (!isnull)
{
/* use user-defined inner consistent method */
procinfo = index_getprocinfo(index, 1, SPGIST_INNER_CONSISTENT_PROC);
FunctionCall2Coll(procinfo,
index->rd_indcollation[0],
PointerGetDatum(&in),
PointerGetDatum(&out));
}
else
{
/* force all children to be visited */
out.nNodes = in.nNodes;
out.nodeNumbers = (int *) palloc(sizeof(int) * in.nNodes);
for (i = 0; i < in.nNodes; i++)
out.nodeNumbers[i] = i;
}
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
MemoryContextSwitchTo(oldCtx);
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
/* If allTheSame, they should all or none of 'em match */
if (innerTuple->allTheSame)
if (out.nNodes != 0 && out.nNodes != in.nNodes)
elog(ERROR, "inconsistent inner_consistent results for allTheSame inner tuple");
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
for (i = 0; i < out.nNodes; i++)
{
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 21:20:04 +02:00
int nodeN = out.nodeNumbers[i];
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
Assert(nodeN >= 0 && nodeN < in.nNodes);
if (ItemPointerIsValid(&nodes[nodeN]->t_tid))
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
{
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
ScanStackEntry *newEntry;
/* Create new work item for this node */
newEntry = palloc(sizeof(ScanStackEntry));
newEntry->ptr = nodes[nodeN]->t_tid;
if (out.levelAdds)
newEntry->level = stackEntry->level + out.levelAdds[i];
else
newEntry->level = stackEntry->level;
/* Must copy value out of temp context */
if (out.reconstructedValues)
newEntry->reconstructedValue =
datumCopy(out.reconstructedValues[i],
so->state.attType.attbyval,
so->state.attType.attlen);
else
newEntry->reconstructedValue = (Datum) 0;
so->scanStack = lcons(newEntry, so->scanStack);
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
}
}
}
/* done with this scan stack entry */
freeScanStackEntry(so, stackEntry);
/* clear temp context before proceeding to the next one */
MemoryContextReset(so->tempCxt);
}
if (buffer != InvalidBuffer)
UnlockReleaseBuffer(buffer);
}
/* storeRes subroutine for getbitmap case */
static void
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
storeBitmap(SpGistScanOpaque so, ItemPointer heapPtr,
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
Datum leafValue, bool isnull, bool recheck)
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
{
tbm_add_tuples(so->tbm, heapPtr, 1, recheck);
so->ntids++;
}
Datum
spggetbitmap(PG_FUNCTION_ARGS)
{
IndexScanDesc scan = (IndexScanDesc) PG_GETARG_POINTER(0);
TIDBitmap *tbm = (TIDBitmap *) PG_GETARG_POINTER(1);
SpGistScanOpaque so = (SpGistScanOpaque) scan->opaque;
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
/* Copy want_itup to *so so we don't need to pass it around separately */
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
so->want_itup = false;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
so->tbm = tbm;
so->ntids = 0;
spgWalk(scan->indexRelation, so, true, storeBitmap);
PG_RETURN_INT64(so->ntids);
}
/* storeRes subroutine for gettuple case */
static void
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
storeGettuple(SpGistScanOpaque so, ItemPointer heapPtr,
Teach SPGiST to store nulls and do whole-index scans. This patch fixes the other major compatibility-breaking limitation of SPGiST, that it didn't store anything for null values of the indexed column, and so could not support whole-index scans or "x IS NULL" tests. The approach is to create a wholly separate search tree for the null entries, and use fixed "allTheSame" insertion and search rules when processing this tree, instead of calling the index opclass methods. This way the opclass methods do not need to worry about dealing with nulls. Catversion bump is for pg_am updates as well as the change in on-disk format of SPGiST indexes; there are some tweaks in SPGiST WAL records as well. Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev. (The original also stored nulls separately, but it reused GIN code to do so; which required undesirable compromises in the on-disk format, and would likely lead to bugs due to the GIN code being required to work in two very different contexts.)
2012-03-11 21:29:04 +01:00
Datum leafValue, bool isnull, bool recheck)
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
{
Assert(so->nPtrs < MaxIndexTuplesPerPage);
so->heapPtrs[so->nPtrs] = *heapPtr;
so->recheck[so->nPtrs] = recheck;
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
if (so->want_itup)
{
/*
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 21:20:04 +02:00
* Reconstruct desired IndexTuple. We have to copy the datum out of
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
* the temp context anyway, so we may as well create the tuple here.
*/
so->indexTups[so->nPtrs] = index_form_tuple(so->indexTupDesc,
&leafValue,
&isnull);
}
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
so->nPtrs++;
}
Datum
spggettuple(PG_FUNCTION_ARGS)
{
IndexScanDesc scan = (IndexScanDesc) PG_GETARG_POINTER(0);
ScanDirection dir = (ScanDirection) PG_GETARG_INT32(1);
SpGistScanOpaque so = (SpGistScanOpaque) scan->opaque;
if (dir != ForwardScanDirection)
elog(ERROR, "SP-GiST only supports forward scan direction");
Restructure SPGiST opclass interface API to support whole-index scans. The original API definition was incapable of supporting whole-index scans because there was no way to invoke leaf-value reconstruction without checking any qual conditions. Also, it was inefficient for multiple-qual-condition scans because value reconstruction got done over again for each qual condition, and because other internal work in the consistent functions likewise had to be done for each qual. To fix these issues, pass the whole scankey array to the opclass consistent functions, instead of only letting them see one item at a time. (Essentially, the loop over scankey entries is now inside the consistent functions not outside them. This makes the consistent functions a bit more complicated, but not unreasonably so.) In itself this commit does nothing except save a few cycles in multiple-qual-condition index scans, since we can't support whole-index scans on SPGiST indexes until nulls are included in the index. However, I consider this a must-fix for 9.2 because once we release it will get very much harder to change the opclass API definition.
2012-03-11 00:36:49 +01:00
/* Copy want_itup to *so so we don't need to pass it around separately */
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
so->want_itup = scan->xs_want_itup;
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
for (;;)
{
if (so->iPtr < so->nPtrs)
{
/* continuing to return tuples from a leaf page */
scan->xs_ctup.t_self = so->heapPtrs[so->iPtr];
scan->xs_recheck = so->recheck[so->iPtr];
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
scan->xs_itup = so->indexTups[so->iPtr];
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
so->iPtr++;
PG_RETURN_BOOL(true);
}
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
if (so->want_itup)
{
/* Must pfree IndexTuples to avoid memory leak */
Run pgindent on 9.2 source tree in preparation for first 9.3 commit-fest.
2012-06-10 21:20:04 +02:00
int i;
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
for (i = 0; i < so->nPtrs; i++)
pfree(so->indexTups[i]);
}
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
so->iPtr = so->nPtrs = 0;
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
Add SP-GiST (space-partitioned GiST) index access method. SP-GiST is comparable to GiST in flexibility, but supports non-balanced partitioned search structures rather than balanced trees. As described at PGCon 2011, this new indexing structure can beat GiST in both index build time and query speed for search problems that it is well matched to. There are a number of areas that could still use improvement, but at this point the code seems committable. Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
2011-12-17 22:41:16 +01:00
spgWalk(scan->indexRelation, so, false, storeGettuple);
if (so->nPtrs == 0)
break; /* must have completed scan */
}
PG_RETURN_BOOL(false);
}
Replace simple constant pg_am.amcanreturn with an AM support function. The need for this was debated when we put in the index-only-scan feature, but at the time we had no near-term expectation of having AMs that could support such scans for only some indexes; so we kept it simple. However, the SP-GiST AM forces the issue, so let's fix it. This patch only installs the new API; no behavior actually changes.
2011-12-18 21:49:00 +01:00
Datum
spgcanreturn(PG_FUNCTION_ARGS)
{
Teach SP-GiST to do index-only scans. Operator classes can specify whether or not they support this; this preserves the flexibility to use lossy representations within an index. In passing, move constant data about a given index into the rd_amcache cache area, instead of doing fresh lookups each time we start an index operation. This is mainly to try to make sure that spgcanreturn() has insignificant cost; I still don't have any proof that it matters for actual index accesses. Also, get rid of useless copying of FmgrInfo pointers; we can perfectly well use the relcache's versions in-place.
2011-12-19 20:58:41 +01:00
Relation index = (Relation) PG_GETARG_POINTER(0);
SpGistCache *cache;
/* We can do it if the opclass config function says so */
cache = spgGetCache(index);
PG_RETURN_BOOL(cache->config.canReturnData);
Replace simple constant pg_am.amcanreturn with an AM support function. The need for this was debated when we put in the index-only-scan feature, but at the time we had no near-term expectation of having AMs that could support such scans for only some indexes; so we kept it simple. However, the SP-GiST AM forces the issue, so let's fix it. This patch only installs the new API; no behavior actually changes.
2011-12-18 21:49:00 +01:00
}