From 03e56f798e365763486b03a2630fbc3190ccd29a Mon Sep 17 00:00:00 2001 From: Tom Lane Date: Sat, 10 Mar 2012 18:36:49 -0500 Subject: [PATCH] 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. --- doc/src/sgml/spgist.sgml | 38 ++- src/backend/access/spgist/spgkdtreeproc.c | 204 +++++------- src/backend/access/spgist/spgquadtreeproc.c | 232 +++++++------ src/backend/access/spgist/spgscan.c | 349 ++++++++++---------- src/backend/access/spgist/spgtextproc.c | 218 ++++++------ src/include/access/spgist.h | 8 +- src/include/access/spgist_private.h | 6 +- 7 files changed, 523 insertions(+), 532 deletions(-) diff --git a/doc/src/sgml/spgist.sgml b/doc/src/sgml/spgist.sgml index dcc3cc2d73..0202dbcdd5 100644 --- a/doc/src/sgml/spgist.sgml +++ b/doc/src/sgml/spgist.sgml @@ -439,8 +439,8 @@ CREATE FUNCTION my_inner_consistent(internal, internal) RETURNS void ... typedef struct spgInnerConsistentIn { - StrategyNumber strategy; /* operator strategy number */ - Datum query; /* operator's RHS value */ + ScanKey scankeys; /* array of operators and comparison values */ + int nkeys; /* length of array */ Datum reconstructedValue; /* value reconstructed at parent */ int level; /* current level (counting from zero) */ @@ -463,8 +463,17 @@ typedef struct spgInnerConsistentOut } spgInnerConsistentOut; - strategy and - query describe the index search condition. + The array scankeys, of length nkeys, + describes the index search condition(s). These conditions are + combined with AND — only index entries that satisfy all of + them are interesting. (Note that nkeys = 0 implies + that all index entries satisfy the query.) Usually the consistent + function only cares about the sk_strategy and + sk_argument fields of each array entry, which + respectively give the indexable operator and comparison value. + In particular it is not necessary to check sk_flags to + see if the comparison value is NULL, because the SP-GiST core code + will filter out such conditions. reconstructedValue is the value reconstructed for the parent tuple; it is (Datum) 0 at the root level or if the inner_consistent function did not provide a value at the @@ -527,8 +536,8 @@ CREATE FUNCTION my_leaf_consistent(internal, internal) RETURNS bool ... typedef struct spgLeafConsistentIn { - StrategyNumber strategy; /* operator strategy number */ - Datum query; /* operator's RHS value */ + ScanKey scankeys; /* array of operators and comparison values */ + int nkeys; /* length of array */ Datum reconstructedValue; /* value reconstructed at parent */ int level; /* current level (counting from zero) */ @@ -544,8 +553,17 @@ typedef struct spgLeafConsistentOut } spgLeafConsistentOut; - strategy and - query define the index search condition. + The array scankeys, of length nkeys, + describes the index search condition(s). These conditions are + combined with AND — only index entries that satisfy all of + them satisfy the query. (Note that nkeys = 0 implies + that all index entries satisfy the query.) Usually the consistent + function only cares about the sk_strategy and + sk_argument fields of each array entry, which + respectively give the indexable operator and comparison value. + In particular it is not necessary to check sk_flags to + see if the comparison value is NULL, because the SP-GiST core code + will filter out such conditions. reconstructedValue is the value reconstructed for the parent tuple; it is (Datum) 0 at the root level or if the inner_consistent function did not provide a value at the @@ -566,8 +584,8 @@ typedef struct spgLeafConsistentOut leafValue must be set to the value originally supplied to be indexed for this leaf tuple. Also, recheck may be set to true if the match - is uncertain and so the operator must be re-applied to the actual heap - tuple to verify the match. + is uncertain and so the operator(s) must be re-applied to the actual + heap tuple to verify the match. diff --git a/src/backend/access/spgist/spgkdtreeproc.c b/src/backend/access/spgist/spgkdtreeproc.c index eca972a6f0..adfe287581 100644 --- a/src/backend/access/spgist/spgkdtreeproc.c +++ b/src/backend/access/spgist/spgkdtreeproc.c @@ -159,11 +159,10 @@ spg_kd_inner_consistent(PG_FUNCTION_ARGS) { spgInnerConsistentIn *in = (spgInnerConsistentIn *) PG_GETARG_POINTER(0); spgInnerConsistentOut *out = (spgInnerConsistentOut *) PG_GETARG_POINTER(1); - Point *query; - BOX *boxQuery; double coord; + int which; + int i; - query = DatumGetPointP(in->query); Assert(in->hasPrefix); coord = DatumGetFloat8(in->prefixDatum); @@ -171,123 +170,96 @@ spg_kd_inner_consistent(PG_FUNCTION_ARGS) elog(ERROR, "allTheSame should not occur for k-d trees"); Assert(in->nNodes == 2); + + /* "which" is a bitmask of children that satisfy all constraints */ + which = (1 << 1) | (1 << 2); + + for (i = 0; i < in->nkeys; i++) + { + Point *query = DatumGetPointP(in->scankeys[i].sk_argument); + BOX *boxQuery; + + switch (in->scankeys[i].sk_strategy) + { + case RTLeftStrategyNumber: + if ((in->level % 2) != 0 && FPlt(query->x, coord)) + which &= (1 << 1); + break; + case RTRightStrategyNumber: + if ((in->level % 2) != 0 && FPgt(query->x, coord)) + which &= (1 << 2); + break; + case RTSameStrategyNumber: + if ((in->level % 2) != 0) + { + if (FPlt(query->x, coord)) + which &= (1 << 1); + else if (FPgt(query->x, coord)) + which &= (1 << 2); + } + else + { + if (FPlt(query->y, coord)) + which &= (1 << 1); + else if (FPgt(query->y, coord)) + which &= (1 << 2); + } + break; + case RTBelowStrategyNumber: + if ((in->level % 2) == 0 && FPlt(query->y, coord)) + which &= (1 << 1); + break; + case RTAboveStrategyNumber: + if ((in->level % 2) == 0 && FPgt(query->y, coord)) + which &= (1 << 2); + break; + case RTContainedByStrategyNumber: + + /* + * For this operator, the query is a box not a point. We + * cheat to the extent of assuming that DatumGetPointP won't + * do anything that would be bad for a pointer-to-box. + */ + boxQuery = DatumGetBoxP(in->scankeys[i].sk_argument); + + if ((in->level % 2) != 0) + { + if (FPlt(boxQuery->high.x, coord)) + which &= (1 << 1); + else if (FPgt(boxQuery->low.x, coord)) + which &= (1 << 2); + } + else + { + if (FPlt(boxQuery->high.y, coord)) + which &= (1 << 1); + else if (FPgt(boxQuery->low.y, coord)) + which &= (1 << 2); + } + break; + default: + elog(ERROR, "unrecognized strategy number: %d", + in->scankeys[i].sk_strategy); + break; + } + + if (which == 0) + break; /* no need to consider remaining conditions */ + } + + /* We must descend into the children identified by which */ out->nodeNumbers = (int *) palloc(sizeof(int) * 2); + out->nNodes = 0; + for (i = 1; i <= 2; i++) + { + if (which & (1 << i)) + out->nodeNumbers[out->nNodes++] = i - 1; + } + + /* Set up level increments, too */ out->levelAdds = (int *) palloc(sizeof(int) * 2); out->levelAdds[0] = 1; out->levelAdds[1] = 1; - out->nNodes = 0; - - switch (in->strategy) - { - case RTLeftStrategyNumber: - out->nNodes = 1; - out->nodeNumbers[0] = 0; - - if ((in->level % 2) == 0 || FPge(query->x, coord)) - { - out->nodeNumbers[1] = 1; - out->nNodes++; - } - break; - case RTRightStrategyNumber: - out->nNodes = 1; - out->nodeNumbers[0] = 1; - - if ((in->level % 2) == 0 || FPle(query->x, coord)) - { - out->nodeNumbers[1] = 0; - out->nNodes++; - } - break; - case RTSameStrategyNumber: - if (in->level % 2) - { - if (FPle(query->x, coord)) - { - out->nodeNumbers[out->nNodes] = 0; - out->nNodes++; - } - if (FPge(query->x, coord)) - { - out->nodeNumbers[out->nNodes] = 1; - out->nNodes++; - } - } - else - { - if (FPle(query->y, coord)) - { - out->nodeNumbers[out->nNodes] = 0; - out->nNodes++; - } - if (FPge(query->y, coord)) - { - out->nodeNumbers[out->nNodes] = 1; - out->nNodes++; - } - } - break; - case RTBelowStrategyNumber: - out->nNodes = 1; - out->nodeNumbers[0] = 0; - - if ((in->level % 2) == 1 || FPge(query->y, coord)) - { - out->nodeNumbers[1] = 1; - out->nNodes++; - } - break; - case RTAboveStrategyNumber: - out->nNodes = 1; - out->nodeNumbers[0] = 1; - - if ((in->level % 2) == 1 || FPle(query->y, coord)) - { - out->nodeNumbers[1] = 0; - out->nNodes++; - } - break; - case RTContainedByStrategyNumber: - - /* - * For this operator, the query is a box not a point. We cheat to - * the extent of assuming that DatumGetPointP won't do anything - * that would be bad for a pointer-to-box. - */ - boxQuery = DatumGetBoxP(in->query); - - out->nNodes = 1; - if (in->level % 2) - { - if (FPlt(boxQuery->high.x, coord)) - out->nodeNumbers[0] = 0; - else if (FPgt(boxQuery->low.x, coord)) - out->nodeNumbers[0] = 1; - else - { - out->nodeNumbers[0] = 0; - out->nodeNumbers[1] = 1; - out->nNodes = 2; - } - } - else - { - if (FPlt(boxQuery->high.y, coord)) - out->nodeNumbers[0] = 0; - else if (FPgt(boxQuery->low.y, coord)) - out->nodeNumbers[0] = 1; - else - { - out->nodeNumbers[0] = 0; - out->nodeNumbers[1] = 1; - out->nNodes = 2; - } - } - break; - default: - elog(ERROR, "unrecognized strategy number: %d", in->strategy); - break; - } PG_RETURN_VOID(); } diff --git a/src/backend/access/spgist/spgquadtreeproc.c b/src/backend/access/spgist/spgquadtreeproc.c index 231749e992..10fafe5864 100644 --- a/src/backend/access/spgist/spgquadtreeproc.c +++ b/src/backend/access/spgist/spgquadtreeproc.c @@ -190,45 +190,21 @@ spg_quad_picksplit(PG_FUNCTION_ARGS) } -/* Subroutine to fill out->nodeNumbers[] for spg_quad_inner_consistent */ -static void -setNodes(spgInnerConsistentOut *out, bool isAll, int first, int second) -{ - if (isAll) - { - out->nNodes = 4; - out->nodeNumbers[0] = 0; - out->nodeNumbers[1] = 1; - out->nodeNumbers[2] = 2; - out->nodeNumbers[3] = 3; - } - else - { - out->nNodes = 2; - out->nodeNumbers[0] = first - 1; - out->nodeNumbers[1] = second - 1; - } -} - - Datum spg_quad_inner_consistent(PG_FUNCTION_ARGS) { spgInnerConsistentIn *in = (spgInnerConsistentIn *) PG_GETARG_POINTER(0); spgInnerConsistentOut *out = (spgInnerConsistentOut *) PG_GETARG_POINTER(1); - Point *query, - *centroid; - BOX *boxQuery; + Point *centroid; + int which; + int i; - query = DatumGetPointP(in->query); Assert(in->hasPrefix); centroid = DatumGetPointP(in->prefixDatum); if (in->allTheSame) { /* Report that all nodes should be visited */ - int i; - out->nNodes = in->nNodes; out->nodeNumbers = (int *) palloc(sizeof(int) * in->nNodes); for (i = 0; i < in->nNodes; i++) @@ -237,76 +213,86 @@ spg_quad_inner_consistent(PG_FUNCTION_ARGS) } Assert(in->nNodes == 4); - out->nodeNumbers = (int *) palloc(sizeof(int) * 4); - switch (in->strategy) + /* "which" is a bitmask of quadrants that satisfy all constraints */ + which = (1 << 1) | (1 << 2) | (1 << 3) | (1 << 4); + + for (i = 0; i < in->nkeys; i++) { - case RTLeftStrategyNumber: - setNodes(out, SPTEST(point_left, centroid, query), 3, 4); - break; - case RTRightStrategyNumber: - setNodes(out, SPTEST(point_right, centroid, query), 1, 2); - break; - case RTSameStrategyNumber: - out->nNodes = 1; - out->nodeNumbers[0] = getQuadrant(centroid, query) - 1; - break; - case RTBelowStrategyNumber: - setNodes(out, SPTEST(point_below, centroid, query), 2, 3); - break; - case RTAboveStrategyNumber: - setNodes(out, SPTEST(point_above, centroid, query), 1, 4); - break; - case RTContainedByStrategyNumber: + Point *query = DatumGetPointP(in->scankeys[i].sk_argument); + BOX *boxQuery; - /* - * For this operator, the query is a box not a point. We cheat to - * the extent of assuming that DatumGetPointP won't do anything - * that would be bad for a pointer-to-box. - */ - boxQuery = DatumGetBoxP(in->query); + switch (in->scankeys[i].sk_strategy) + { + case RTLeftStrategyNumber: + if (SPTEST(point_right, centroid, query)) + which &= (1 << 3) | (1 << 4); + break; + case RTRightStrategyNumber: + if (SPTEST(point_left, centroid, query)) + which &= (1 << 1) | (1 << 2); + break; + case RTSameStrategyNumber: + which &= (1 << getQuadrant(centroid, query)); + break; + case RTBelowStrategyNumber: + if (SPTEST(point_above, centroid, query)) + which &= (1 << 2) | (1 << 3); + break; + case RTAboveStrategyNumber: + if (SPTEST(point_below, centroid, query)) + which &= (1 << 1) | (1 << 4); + break; + case RTContainedByStrategyNumber: - if (DatumGetBool(DirectFunctionCall2(box_contain_pt, - PointerGetDatum(boxQuery), - PointerGetDatum(centroid)))) - { - /* centroid is in box, so descend to all quadrants */ - setNodes(out, true, 0, 0); - } - else - { - /* identify quadrant(s) containing all corners of box */ - Point p; - int i, - r = 0; + /* + * For this operator, the query is a box not a point. We + * cheat to the extent of assuming that DatumGetPointP won't + * do anything that would be bad for a pointer-to-box. + */ + boxQuery = DatumGetBoxP(in->scankeys[i].sk_argument); - p = boxQuery->low; - r |= 1 << (getQuadrant(centroid, &p) - 1); - - p.y = boxQuery->high.y; - r |= 1 << (getQuadrant(centroid, &p) - 1); - - p = boxQuery->high; - r |= 1 << (getQuadrant(centroid, &p) - 1); - - p.x = boxQuery->low.x; - r |= 1 << (getQuadrant(centroid, &p) - 1); - - /* we must descend into those quadrant(s) */ - out->nNodes = 0; - for (i = 0; i < 4; i++) + if (DatumGetBool(DirectFunctionCall2(box_contain_pt, + PointerGetDatum(boxQuery), + PointerGetDatum(centroid)))) { - if (r & (1 << i)) - { - out->nodeNumbers[out->nNodes] = i; - out->nNodes++; - } + /* centroid is in box, so all quadrants are OK */ } - } - break; - default: - elog(ERROR, "unrecognized strategy number: %d", in->strategy); - break; + else + { + /* identify quadrant(s) containing all corners of box */ + Point p; + int r = 0; + + p = boxQuery->low; + r |= 1 << getQuadrant(centroid, &p); + p.y = boxQuery->high.y; + r |= 1 << getQuadrant(centroid, &p); + p = boxQuery->high; + r |= 1 << getQuadrant(centroid, &p); + p.x = boxQuery->low.x; + r |= 1 << getQuadrant(centroid, &p); + + which &= r; + } + break; + default: + elog(ERROR, "unrecognized strategy number: %d", + in->scankeys[i].sk_strategy); + break; + } + + if (which == 0) + break; /* no need to consider remaining conditions */ + } + + /* We must descend into the quadrant(s) identified by which */ + out->nodeNumbers = (int *) palloc(sizeof(int) * 4); + out->nNodes = 0; + for (i = 1; i <= 4; i++) + { + if (which & (1 << i)) + out->nodeNumbers[out->nNodes++] = i - 1; } PG_RETURN_VOID(); @@ -318,9 +304,9 @@ spg_quad_leaf_consistent(PG_FUNCTION_ARGS) { spgLeafConsistentIn *in = (spgLeafConsistentIn *) PG_GETARG_POINTER(0); spgLeafConsistentOut *out = (spgLeafConsistentOut *) PG_GETARG_POINTER(1); - Point *query = DatumGetPointP(in->query); Point *datum = DatumGetPointP(in->leafDatum); bool res; + int i; /* all tests are exact */ out->recheck = false; @@ -328,35 +314,45 @@ spg_quad_leaf_consistent(PG_FUNCTION_ARGS) /* leafDatum is what it is... */ out->leafValue = in->leafDatum; - switch (in->strategy) + /* Perform the required comparison(s) */ + res = true; + for (i = 0; i < in->nkeys; i++) { - case RTLeftStrategyNumber: - res = SPTEST(point_left, datum, query); - break; - case RTRightStrategyNumber: - res = SPTEST(point_right, datum, query); - break; - case RTSameStrategyNumber: - res = SPTEST(point_eq, datum, query); - break; - case RTBelowStrategyNumber: - res = SPTEST(point_below, datum, query); - break; - case RTAboveStrategyNumber: - res = SPTEST(point_above, datum, query); - break; - case RTContainedByStrategyNumber: + Point *query = DatumGetPointP(in->scankeys[i].sk_argument); - /* - * For this operator, the query is a box not a point. We cheat to - * the extent of assuming that DatumGetPointP won't do anything - * that would be bad for a pointer-to-box. - */ - res = SPTEST(box_contain_pt, query, datum); - break; - default: - elog(ERROR, "unrecognized strategy number: %d", in->strategy); - res = false; + switch (in->scankeys[i].sk_strategy) + { + case RTLeftStrategyNumber: + res = SPTEST(point_left, datum, query); + break; + case RTRightStrategyNumber: + res = SPTEST(point_right, datum, query); + break; + case RTSameStrategyNumber: + res = SPTEST(point_eq, datum, query); + break; + case RTBelowStrategyNumber: + res = SPTEST(point_below, datum, query); + break; + case RTAboveStrategyNumber: + res = SPTEST(point_above, datum, query); + break; + case RTContainedByStrategyNumber: + + /* + * For this operator, the query is a box not a point. We + * cheat to the extent of assuming that DatumGetPointP won't + * do anything that would be bad for a pointer-to-box. + */ + res = SPTEST(box_contain_pt, query, datum); + break; + default: + elog(ERROR, "unrecognized strategy number: %d", + in->scankeys[i].sk_strategy); + break; + } + + if (!res) break; } diff --git a/src/backend/access/spgist/spgscan.c b/src/backend/access/spgist/spgscan.c index 22cfcc8792..99b0852611 100644 --- a/src/backend/access/spgist/spgscan.c +++ b/src/backend/access/spgist/spgscan.c @@ -56,18 +56,25 @@ freeScanStack(SpGistScanOpaque so) } /* - * Initialize scanStack with a single entry for the root page, resetting + * Initialize scanStack to search the root page, resetting * any previously active scan */ static void resetSpGistScanOpaque(SpGistScanOpaque so) { - ScanStackEntry *startEntry = palloc0(sizeof(ScanStackEntry)); - - ItemPointerSet(&startEntry->ptr, SPGIST_HEAD_BLKNO, FirstOffsetNumber); + ScanStackEntry *startEntry; freeScanStack(so); - so->scanStack = list_make1(startEntry); + + Assert(!so->searchNulls); /* XXX fixme */ + + if (so->searchNonNulls) + { + /* Stack a work item to scan the non-null index entries */ + startEntry = (ScanStackEntry *) palloc0(sizeof(ScanStackEntry)); + ItemPointerSet(&startEntry->ptr, SPGIST_HEAD_BLKNO, FirstOffsetNumber); + so->scanStack = list_make1(startEntry); + } if (so->want_itup) { @@ -80,6 +87,82 @@ resetSpGistScanOpaque(SpGistScanOpaque so) so->iPtr = so->nPtrs = 0; } +/* + * 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 + * opclass consistent functions need to deal with. We assume all SPGiST- + * 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; + } +} + Datum spgbeginscan(PG_FUNCTION_ARGS) { @@ -92,13 +175,16 @@ spgbeginscan(PG_FUNCTION_ARGS) scan = RelationGetIndexScan(rel, keysz, 0); so = (SpGistScanOpaque) palloc0(sizeof(SpGistScanOpaqueData)); + if (keysz > 0) + so->keyData = (ScanKey) palloc(sizeof(ScanKeyData) * keysz); + else + so->keyData = NULL; initSpGistState(&so->state, scan->indexRelation); so->tempCxt = AllocSetContextCreate(CurrentMemoryContext, "SP-GiST search temporary context", ALLOCSET_DEFAULT_MINSIZE, ALLOCSET_DEFAULT_INITSIZE, ALLOCSET_DEFAULT_MAXSIZE); - resetSpGistScanOpaque(so); /* Set up indexTupDesc and xs_itupdesc in case it's an index-only scan */ so->indexTupDesc = scan->xs_itupdesc = RelationGetDescr(rel); @@ -115,12 +201,17 @@ spgrescan(PG_FUNCTION_ARGS) SpGistScanOpaque so = (SpGistScanOpaque) scan->opaque; ScanKey scankey = (ScanKey) PG_GETARG_POINTER(1); + /* copy scankeys into local storage */ if (scankey && scan->numberOfKeys > 0) { memmove(scan->keyData, scankey, scan->numberOfKeys * sizeof(ScanKeyData)); } + /* preprocess scankeys, set up the representation in *so */ + spgPrepareScanKeys(scan); + + /* set up starting stack entries */ resetSpGistScanOpaque(so); PG_RETURN_VOID(); @@ -162,53 +253,34 @@ spgLeafTest(Relation index, SpGistScanOpaque so, Datum leafDatum, int level, Datum reconstructedValue, Datum *leafValue, bool *recheck) { - bool result = true; + bool result; spgLeafConsistentIn in; spgLeafConsistentOut out; FmgrInfo *procinfo; MemoryContext oldCtx; - int i; - *leafValue = (Datum) 0; - *recheck = false; + /* use temp context for calling leaf_consistent */ + oldCtx = MemoryContextSwitchTo(so->tempCxt); - /* set up values that are the same for all quals */ + in.scankeys = so->keyData; + in.nkeys = so->numberOfKeys; in.reconstructedValue = reconstructedValue; in.level = level; in.returnData = so->want_itup; in.leafDatum = leafDatum; - /* Apply each leaf consistency check, working in the temp context */ - oldCtx = MemoryContextSwitchTo(so->tempCxt); + out.leafValue = (Datum) 0; + out.recheck = false; procinfo = index_getprocinfo(index, 1, SPGIST_LEAF_CONSISTENT_PROC); + result = DatumGetBool(FunctionCall2Coll(procinfo, + index->rd_indcollation[0], + PointerGetDatum(&in), + PointerGetDatum(&out))); - for (i = 0; i < so->numberOfKeys; i++) - { - ScanKey skey = &so->keyData[i]; + *leafValue = out.leafValue; + *recheck = out.recheck; - /* Assume SPGiST-indexable operators are strict */ - if (skey->sk_flags & SK_ISNULL) - { - result = false; - break; - } - - in.strategy = skey->sk_strategy; - in.query = skey->sk_argument; - - out.leafValue = (Datum) 0; - out.recheck = false; - - result = DatumGetBool(FunctionCall2Coll(procinfo, - skey->sk_collation, - PointerGetDatum(&in), - PointerGetDatum(&out))); - *leafValue = out.leafValue; - *recheck |= out.recheck; - if (!result) - break; - } MemoryContextSwitchTo(oldCtx); return result; @@ -349,8 +421,13 @@ redirect: else /* page is inner */ { SpGistInnerTuple innerTuple; + spgInnerConsistentIn in; + spgInnerConsistentOut out; + FmgrInfo *procinfo; + SpGistNodeTuple *nodes; SpGistNodeTuple node; int i; + MemoryContext oldCtx; innerTuple = (SpGistInnerTuple) PageGetItem(page, PageGetItemId(page, offset)); @@ -368,144 +445,68 @@ redirect: innerTuple->tupstate); } - if (so->numberOfKeys == 0) + /* 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) { - /* - * This case cannot happen at the moment, because we don't - * set pg_am.amoptionalkey for SP-GiST. In order for full - * index scans to produce correct answers, we'd need to - * index nulls, which we don't. - */ - Assert(false); - -#ifdef NOT_USED - /* - * A full index scan could be done approximately like this, - * but note that reconstruction of indexed values would be - * impossible unless the API for inner_consistent is changed. - */ - SGITITERATE(innerTuple, i, node) - { - if (ItemPointerIsValid(&node->t_tid)) - { - ScanStackEntry *newEntry = palloc(sizeof(ScanStackEntry)); - - newEntry->ptr = node->t_tid; - newEntry->level = -1; - newEntry->reconstructedValue = (Datum) 0; - so->scanStack = lcons(newEntry, so->scanStack); - } - } -#endif + nodes[i] = node; } - else + + memset(&out, 0, sizeof(out)); + + procinfo = index_getprocinfo(index, 1, SPGIST_INNER_CONSISTENT_PROC); + FunctionCall2Coll(procinfo, + index->rd_indcollation[0], + PointerGetDatum(&in), + PointerGetDatum(&out)); + + MemoryContextSwitchTo(oldCtx); + + /* 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"); + + for (i = 0; i < out.nNodes; i++) { - spgInnerConsistentIn in; - spgInnerConsistentOut out; - FmgrInfo *procinfo; - SpGistNodeTuple *nodes; - int *andMap; - int *levelAdds; - Datum *reconstructedValues; - int j, - nMatches = 0; - MemoryContext oldCtx; + int nodeN = out.nodeNumbers[i]; - /* use temp context for calling inner_consistent */ - oldCtx = MemoryContextSwitchTo(so->tempCxt); - - /* set up values that are the same for all scankeys */ - 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) + Assert(nodeN >= 0 && nodeN < in.nNodes); + if (ItemPointerIsValid(&nodes[nodeN]->t_tid)) { - nodes[i] = node; - } + ScanStackEntry *newEntry; - andMap = (int *) palloc0(sizeof(int) * in.nNodes); - levelAdds = (int *) palloc0(sizeof(int) * in.nNodes); - reconstructedValues = (Datum *) palloc0(sizeof(Datum) * in.nNodes); + /* 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; - procinfo = index_getprocinfo(index, 1, SPGIST_INNER_CONSISTENT_PROC); - - for (j = 0; j < so->numberOfKeys; j++) - { - ScanKey skey = &so->keyData[j]; - - /* Assume SPGiST-indexable operators are strict */ - if (skey->sk_flags & SK_ISNULL) - { - nMatches = 0; - break; - } - - in.strategy = skey->sk_strategy; - in.query = skey->sk_argument; - - memset(&out, 0, sizeof(out)); - - FunctionCall2Coll(procinfo, - skey->sk_collation, - PointerGetDatum(&in), - PointerGetDatum(&out)); - - /* 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"); - - nMatches = 0; - for (i = 0; i < out.nNodes; i++) - { - int nodeN = out.nodeNumbers[i]; - - andMap[nodeN]++; - if (andMap[nodeN] == j + 1) - nMatches++; - if (out.levelAdds) - levelAdds[nodeN] = out.levelAdds[i]; - if (out.reconstructedValues) - reconstructedValues[nodeN] = out.reconstructedValues[i]; - } - - /* quit as soon as all nodes have failed some qual */ - if (nMatches == 0) - break; - } - - MemoryContextSwitchTo(oldCtx); - - if (nMatches > 0) - { - for (i = 0; i < in.nNodes; i++) - { - if (andMap[i] == so->numberOfKeys && - ItemPointerIsValid(&nodes[i]->t_tid)) - { - ScanStackEntry *newEntry; - - /* Create new work item for this node */ - newEntry = palloc(sizeof(ScanStackEntry)); - newEntry->ptr = nodes[i]->t_tid; - newEntry->level = stackEntry->level + levelAdds[i]; - /* Must copy value out of temp context */ - newEntry->reconstructedValue = - datumCopy(reconstructedValues[i], - so->state.attType.attbyval, - so->state.attType.attlen); - - so->scanStack = lcons(newEntry, so->scanStack); - } - } + so->scanStack = lcons(newEntry, so->scanStack); } } } @@ -536,10 +537,7 @@ spggetbitmap(PG_FUNCTION_ARGS) TIDBitmap *tbm = (TIDBitmap *) PG_GETARG_POINTER(1); SpGistScanOpaque so = (SpGistScanOpaque) scan->opaque; - /* Copy scankey to *so so we don't need to pass it around separately */ - so->numberOfKeys = scan->numberOfKeys; - so->keyData = scan->keyData; - /* Ditto for the want_itup flag */ + /* Copy want_itup to *so so we don't need to pass it around separately */ so->want_itup = false; so->tbm = tbm; @@ -583,10 +581,7 @@ spggettuple(PG_FUNCTION_ARGS) if (dir != ForwardScanDirection) elog(ERROR, "SP-GiST only supports forward scan direction"); - /* Copy scankey to *so so we don't need to pass it around separately */ - so->numberOfKeys = scan->numberOfKeys; - so->keyData = scan->keyData; - /* Ditto for the want_itup flag */ + /* Copy want_itup to *so so we don't need to pass it around separately */ so->want_itup = scan->xs_want_itup; for (;;) diff --git a/src/backend/access/spgist/spgtextproc.c b/src/backend/access/spgist/spgtextproc.c index b194fc1b13..656015ea7e 100644 --- a/src/backend/access/spgist/spgtextproc.c +++ b/src/backend/access/spgist/spgtextproc.c @@ -362,25 +362,12 @@ spg_text_inner_consistent(PG_FUNCTION_ARGS) { spgInnerConsistentIn *in = (spgInnerConsistentIn *) PG_GETARG_POINTER(0); spgInnerConsistentOut *out = (spgInnerConsistentOut *) PG_GETARG_POINTER(1); - StrategyNumber strategy = in->strategy; - text *inText; - int inSize; - int i; + bool collate_is_c = lc_collate_is_c(PG_GET_COLLATION()); text *reconstrText = NULL; int maxReconstrLen = 0; text *prefixText = NULL; int prefixSize = 0; - - /* - * If it's a collation-aware operator, but the collation is C, we can - * treat it as non-collation-aware. - */ - if (strategy > 10 && - lc_collate_is_c(PG_GET_COLLATION())) - strategy -= 10; - - inText = DatumGetTextPP(in->query); - inSize = VARSIZE_ANY_EXHDR(inText); + int i; /* * Reconstruct values represented at this tuple, including parent data, @@ -431,8 +418,8 @@ spg_text_inner_consistent(PG_FUNCTION_ARGS) { uint8 nodeChar = DatumGetUInt8(in->nodeLabels[i]); int thisLen; - int r; - bool res = false; + bool res = true; + int j; /* If nodeChar is zero, don't include it in data */ if (nodeChar == '\0') @@ -443,38 +430,57 @@ spg_text_inner_consistent(PG_FUNCTION_ARGS) thisLen = maxReconstrLen; } - r = memcmp(VARDATA(reconstrText), VARDATA_ANY(inText), - Min(inSize, thisLen)); - - switch (strategy) + for (j = 0; j < in->nkeys; j++) { - case BTLessStrategyNumber: - case BTLessEqualStrategyNumber: - if (r <= 0) - res = true; - break; - case BTEqualStrategyNumber: - if (r == 0 && inSize >= thisLen) - res = true; - break; - case BTGreaterEqualStrategyNumber: - case BTGreaterStrategyNumber: - if (r >= 0) - res = true; - break; - case BTLessStrategyNumber + 10: - case BTLessEqualStrategyNumber + 10: - case BTGreaterEqualStrategyNumber + 10: - case BTGreaterStrategyNumber + 10: - /* - * with non-C collation we need to traverse whole tree :-( - */ - res = true; - break; - default: - elog(ERROR, "unrecognized strategy number: %d", - in->strategy); - break; + StrategyNumber strategy = in->scankeys[j].sk_strategy; + text *inText; + int inSize; + int r; + + /* + * If it's a collation-aware operator, but the collation is C, we + * can treat it as non-collation-aware. With non-C collation we + * need to traverse whole tree :-( so there's no point in making + * any check here. + */ + if (strategy > 10) + { + if (collate_is_c) + strategy -= 10; + else + continue; + } + + inText = DatumGetTextPP(in->scankeys[j].sk_argument); + inSize = VARSIZE_ANY_EXHDR(inText); + + r = memcmp(VARDATA(reconstrText), VARDATA_ANY(inText), + Min(inSize, thisLen)); + + switch (strategy) + { + case BTLessStrategyNumber: + case BTLessEqualStrategyNumber: + if (r > 0) + res = false; + break; + case BTEqualStrategyNumber: + if (r != 0 || inSize < thisLen) + res = false; + break; + case BTGreaterEqualStrategyNumber: + case BTGreaterStrategyNumber: + if (r < 0) + res = false; + break; + default: + elog(ERROR, "unrecognized strategy number: %d", + in->scankeys[j].sk_strategy); + break; + } + + if (!res) + break; /* no need to consider remaining conditions */ } if (res) @@ -496,16 +502,13 @@ spg_text_leaf_consistent(PG_FUNCTION_ARGS) { spgLeafConsistentIn *in = (spgLeafConsistentIn *) PG_GETARG_POINTER(0); spgLeafConsistentOut *out = (spgLeafConsistentOut *) PG_GETARG_POINTER(1); - StrategyNumber strategy = in->strategy; - text *query = DatumGetTextPP(in->query); int level = in->level; text *leafValue, *reconstrValue = NULL; char *fullValue; int fullLen; - int queryLen; - int r; bool res; + int j; /* all tests are exact */ out->recheck = false; @@ -518,18 +521,8 @@ spg_text_leaf_consistent(PG_FUNCTION_ARGS) Assert(level == 0 ? reconstrValue == NULL : VARSIZE_ANY_EXHDR(reconstrValue) == level); + /* Reconstruct the full string represented by this leaf tuple */ fullLen = level + VARSIZE_ANY_EXHDR(leafValue); - - queryLen = VARSIZE_ANY_EXHDR(query); - - /* - * For an equality check, we needn't reconstruct fullValue if not same - * length; it can't match - */ - if (strategy == BTEqualStrategyNumber && queryLen != fullLen) - PG_RETURN_BOOL(false); - - /* Else, reconstruct the full string represented by this leaf tuple */ if (VARSIZE_ANY_EXHDR(leafValue) == 0 && level > 0) { fullValue = VARDATA(reconstrValue); @@ -549,54 +542,67 @@ spg_text_leaf_consistent(PG_FUNCTION_ARGS) out->leafValue = PointerGetDatum(fullText); } - /* Run the appropriate type of comparison */ - if (strategy > 10) + /* Perform the required comparison(s) */ + res = true; + for (j = 0; j < in->nkeys; j++) { - /* Collation-aware comparison */ - strategy -= 10; + StrategyNumber strategy = in->scankeys[j].sk_strategy; + text *query = DatumGetTextPP(in->scankeys[j].sk_argument); + int queryLen = VARSIZE_ANY_EXHDR(query); + int r; - /* If asserts are enabled, verify encoding of reconstructed string */ - Assert(pg_verifymbstr(fullValue, fullLen, false)); + if (strategy > 10) + { + /* Collation-aware comparison */ + strategy -= 10; - r = varstr_cmp(fullValue, Min(queryLen, fullLen), - VARDATA_ANY(query), Min(queryLen, fullLen), - PG_GET_COLLATION()); - } - else - { - /* Non-collation-aware comparison */ - r = memcmp(fullValue, VARDATA_ANY(query), Min(queryLen, fullLen)); - } + /* If asserts enabled, verify encoding of reconstructed string */ + Assert(pg_verifymbstr(fullValue, fullLen, false)); - if (r == 0) - { - if (queryLen > fullLen) - r = -1; - else if (queryLen < fullLen) - r = 1; - } + r = varstr_cmp(fullValue, Min(queryLen, fullLen), + VARDATA_ANY(query), Min(queryLen, fullLen), + PG_GET_COLLATION()); + } + else + { + /* Non-collation-aware comparison */ + r = memcmp(fullValue, VARDATA_ANY(query), Min(queryLen, fullLen)); + } - switch (strategy) - { - case BTLessStrategyNumber: - res = (r < 0); - break; - case BTLessEqualStrategyNumber: - res = (r <= 0); - break; - case BTEqualStrategyNumber: - res = (r == 0); - break; - case BTGreaterEqualStrategyNumber: - res = (r >= 0); - break; - case BTGreaterStrategyNumber: - res = (r > 0); - break; - default: - elog(ERROR, "unrecognized strategy number: %d", in->strategy); - res = false; - break; + if (r == 0) + { + if (queryLen > fullLen) + r = -1; + else if (queryLen < fullLen) + r = 1; + } + + switch (strategy) + { + case BTLessStrategyNumber: + res = (r < 0); + break; + case BTLessEqualStrategyNumber: + res = (r <= 0); + break; + case BTEqualStrategyNumber: + res = (r == 0); + break; + case BTGreaterEqualStrategyNumber: + res = (r >= 0); + break; + case BTGreaterStrategyNumber: + res = (r > 0); + break; + default: + elog(ERROR, "unrecognized strategy number: %d", + in->scankeys[j].sk_strategy); + res = false; + break; + } + + if (!res) + break; /* no need to consider remaining conditions */ } PG_RETURN_BOOL(res); diff --git a/src/include/access/spgist.h b/src/include/access/spgist.h index cd6de2c98d..8d0205e691 100644 --- a/src/include/access/spgist.h +++ b/src/include/access/spgist.h @@ -128,8 +128,8 @@ typedef struct spgPickSplitOut */ typedef struct spgInnerConsistentIn { - StrategyNumber strategy; /* operator strategy number */ - Datum query; /* operator's RHS value */ + ScanKey scankeys; /* array of operators and comparison values */ + int nkeys; /* length of array */ Datum reconstructedValue; /* value reconstructed at parent */ int level; /* current level (counting from zero) */ @@ -156,8 +156,8 @@ typedef struct spgInnerConsistentOut */ typedef struct spgLeafConsistentIn { - StrategyNumber strategy; /* operator strategy number */ - Datum query; /* operator's RHS value */ + ScanKey scankeys; /* array of operators and comparison values */ + int nkeys; /* length of array */ Datum reconstructedValue; /* value reconstructed at parent */ int level; /* current level (counting from zero) */ diff --git a/src/include/access/spgist_private.h b/src/include/access/spgist_private.h index fa23acf6cd..76ea5a1578 100644 --- a/src/include/access/spgist_private.h +++ b/src/include/access/spgist_private.h @@ -126,7 +126,11 @@ typedef struct SpGistScanOpaqueData SpGistState state; /* see above */ MemoryContext tempCxt; /* short-lived memory context */ - /* Index quals for scan (copied from IndexScanDesc for convenience) */ + /* Control flags showing whether to search nulls and/or non-nulls */ + bool searchNulls; /* scan matches (all) null entries */ + bool searchNonNulls; /* scan matches (some) non-null entries */ + + /* Index quals to be passed to opclass (null-related quals removed) */ int numberOfKeys; /* number of index qualifier conditions */ ScanKey keyData; /* array of index qualifier descriptors */