Changeset: 2d5ddc62f9e9 for MonetDB
URL: https://dev.monetdb.org/hg/MonetDB/rev/2d5ddc62f9e9
Modified Files:
clients/Tests/exports.stable.out
Branch: default
Log Message:
Merge groupjoin branch into default.
diffs (truncated from 5997 to 300 lines):
diff --git a/clients/Tests/MAL-signatures-hge.test
b/clients/Tests/MAL-signatures-hge.test
--- a/clients/Tests/MAL-signatures-hge.test
+++ b/clients/Tests/MAL-signatures-hge.test
@@ -3494,6 +3494,26 @@ command algebra.likeselect(X_0:bat[:str]
PCRElikeselect;
Select all head values of the first input BAT for which the@tail value is
"like" the given (SQL-style) pattern and for@which the head value occurs in the
tail of the second input@BAT.@Input is a dense-headed BAT, output is a
dense-headed BAT with in@the tail the head value of the input BAT for which
the@relationship holds. The output BAT is sorted on the tail value.
algebra
+markjoin
+command algebra.markjoin(X_0:bat[:any_1], X_1:bat[:any_1], X_2:bat[:oid],
X_3:bat[:oid], X_4:lng) (X_5:bat[:oid], X_6:bat[:bit])
+ALGmark2join;
+Mark join with candidate lists
+algebra
+markjoin
+command algebra.markjoin(X_0:bat[:any_1], X_1:bat[:any_1], X_2:bat[:oid],
X_3:bat[:oid], X_4:lng) (X_5:bat[:oid], X_6:bat[:oid], X_7:bat[:bit])
+ALGmark3join;
+Mark join with candidate lists
+algebra
+markjoin
+command algebra.markjoin(X_0:bat[:any_1], X_1:bat[:any_1], X_2:bat[:oid],
X_3:bat[:oid], X_4:bit, X_5:lng) (X_6:bat[:oid], X_7:bat[:bit])
+ALGmarkjoin;
+Left mark join with candidate lists, produces left output and mark flag;
+algebra
+markselect
+command algebra.markselect(X_0:bat[:oid], X_1:bat[:bit], X_2:bat[:bit],
X_3:bit) (X_4:bat[:oid], X_5:bat[:bit])
+ALGmarkselect;
+Group on group-ids, return aggregated anyequal or allnotequal
+algebra
not_like
command algebra.not_like(X_0:str, X_1:str, X_2:str, X_3:bit):bit
PCREnotlike;
@@ -3504,6 +3524,11 @@ command algebra.orderidx(X_0:bat[:any_1]
OIDXorderidx;
Create an order index
algebra
+outercrossproduct
+command algebra.outercrossproduct(X_0:bat[:any_1], X_1:bat[:any_2],
X_2:bat[:oid], X_3:bat[:oid], X_4:bit) (X_5:bat[:oid], X_6:bat[:oid])
+ALGoutercrossproduct3;
+Compute the outer cross product of both input bats
+algebra
outerjoin
command algebra.outerjoin(X_0:bat[:any_1], X_1:bat[:any_1], X_2:bat[:oid],
X_3:bat[:oid], X_4:bit, X_5:bit, X_6:lng):bat[:oid]
ALGouterjoin1;
@@ -3514,6 +3539,11 @@ command algebra.outerjoin(X_0:bat[:any_1
ALGouterjoin;
Left outer join with candidate lists
algebra
+outerselect
+command algebra.outerselect(X_0:bat[:oid], X_1:bat[:bit], X_2:bat[:bit],
X_3:bit) (X_4:bat[:oid], X_5:bat[:bit])
+ALGouterselect;
+Per input lid return atleast one row, if none of the predicates (p) hold,
return a nil, else 'all' true cases.
+algebra
project
pattern algebra.project(X_0:bat[:any_1], X_1:any_3):bat[:any_3]
ALGprojecttail;
diff --git a/clients/Tests/MAL-signatures.test
b/clients/Tests/MAL-signatures.test
--- a/clients/Tests/MAL-signatures.test
+++ b/clients/Tests/MAL-signatures.test
@@ -2929,6 +2929,26 @@ command algebra.likeselect(X_0:bat[:str]
PCRElikeselect;
Select all head values of the first input BAT for which the@tail value is
"like" the given (SQL-style) pattern and for@which the head value occurs in the
tail of the second input@BAT.@Input is a dense-headed BAT, output is a
dense-headed BAT with in@the tail the head value of the input BAT for which
the@relationship holds. The output BAT is sorted on the tail value.
algebra
+markjoin
+command algebra.markjoin(X_0:bat[:any_1], X_1:bat[:any_1], X_2:bat[:oid],
X_3:bat[:oid], X_4:lng) (X_5:bat[:oid], X_6:bat[:bit])
+ALGmark2join;
+Mark join with candidate lists
+algebra
+markjoin
+command algebra.markjoin(X_0:bat[:any_1], X_1:bat[:any_1], X_2:bat[:oid],
X_3:bat[:oid], X_4:lng) (X_5:bat[:oid], X_6:bat[:oid], X_7:bat[:bit])
+ALGmark3join;
+Mark join with candidate lists
+algebra
+markjoin
+command algebra.markjoin(X_0:bat[:any_1], X_1:bat[:any_1], X_2:bat[:oid],
X_3:bat[:oid], X_4:bit, X_5:lng) (X_6:bat[:oid], X_7:bat[:bit])
+ALGmarkjoin;
+Left mark join with candidate lists, produces left output and mark flag;
+algebra
+markselect
+command algebra.markselect(X_0:bat[:oid], X_1:bat[:bit], X_2:bat[:bit],
X_3:bit) (X_4:bat[:oid], X_5:bat[:bit])
+ALGmarkselect;
+Group on group-ids, return aggregated anyequal or allnotequal
+algebra
not_like
command algebra.not_like(X_0:str, X_1:str, X_2:str, X_3:bit):bit
PCREnotlike;
@@ -2939,6 +2959,11 @@ command algebra.orderidx(X_0:bat[:any_1]
OIDXorderidx;
Create an order index
algebra
+outercrossproduct
+command algebra.outercrossproduct(X_0:bat[:any_1], X_1:bat[:any_2],
X_2:bat[:oid], X_3:bat[:oid], X_4:bit) (X_5:bat[:oid], X_6:bat[:oid])
+ALGoutercrossproduct3;
+Compute the outer cross product of both input bats
+algebra
outerjoin
command algebra.outerjoin(X_0:bat[:any_1], X_1:bat[:any_1], X_2:bat[:oid],
X_3:bat[:oid], X_4:bit, X_5:bit, X_6:lng):bat[:oid]
ALGouterjoin1;
@@ -2949,6 +2974,11 @@ command algebra.outerjoin(X_0:bat[:any_1
ALGouterjoin;
Left outer join with candidate lists
algebra
+outerselect
+command algebra.outerselect(X_0:bat[:oid], X_1:bat[:bit], X_2:bat[:bit],
X_3:bit) (X_4:bat[:oid], X_5:bat[:bit])
+ALGouterselect;
+Per input lid return atleast one row, if none of the predicates (p) hold,
return a nil, else 'all' true cases.
+algebra
project
pattern algebra.project(X_0:bat[:any_1], X_1:any_3):bat[:any_3]
ALGprojecttail;
diff --git a/clients/Tests/exports.stable.out b/clients/Tests/exports.stable.out
--- a/clients/Tests/exports.stable.out
+++ b/clients/Tests/exports.stable.out
@@ -159,6 +159,7 @@ BAT *BATintersectcand(BAT *a, BAT *b);
gdk_return BATjoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT *sl, BAT *sr,
bool nil_matches, BUN estimate) __attribute__((__warn_unused_result__));
gdk_return BATkey(BAT *b, bool onoff);
gdk_return BATleftjoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT *sl, BAT *sr,
bool nil_matches, BUN estimate) __attribute__((__warn_unused_result__));
+gdk_return BATmarkjoin(BAT **r1p, BAT **r2p, BAT **r3p, BAT *l, BAT *r, BAT
*sl, BAT *sr, BUN estimate) __attribute__((__warn_unused_result__));
BAT *BATmaskedcands(oid hseq, BUN nr, BAT *masked, bool selected);
void *BATmax(BAT *b, void *aggr);
void *BATmax_skipnil(BAT *b, void *aggr, bit skipnil);
@@ -172,6 +173,7 @@ BAT *BATnil_grp(BAT *l, BAT *g, BAT *e,
bool BATordered(BAT *b);
bool BATordered_rev(BAT *b);
gdk_return BATorderidx(BAT *b, bool stable);
+gdk_return BAToutercross(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT *sl, BAT
*sr, bool max_one) __attribute__((__warn_unused_result__));
gdk_return BATouterjoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT *sl, BAT
*sr, bool nil_matches, bool match_one, BUN estimate)
__attribute__((__warn_unused_result__));
gdk_return BATprint(stream *s, BAT *b);
gdk_return BATprintcolumns(stream *s, int argc, BAT *argv[]);
@@ -1037,6 +1039,8 @@ const char *mal_version(void);
stream *maleventstream;
const char *manifoldRef;
const char *mapiRef;
+const char *markjoinRef;
+const char *markselectRef;
const char *maskRef;
const char *matRef;
const char *maxRef;
@@ -1112,7 +1116,9 @@ str operatorName(int i);
void opt_pipes_reset(void);
str optimizeMALBlock(Client cntxt, MalBlkPtr mb);
const char *optimizerRef;
+const char *outercrossRef;
const char *outerjoinRef;
+const char *outerselectRef;
const char *packIncrementRef;
const char *packRef;
const char *parametersRef;
diff --git a/gdk/gdk.h b/gdk/gdk.h
--- a/gdk/gdk.h
+++ b/gdk/gdk.h
@@ -2302,9 +2302,13 @@ gdk_export BAT *BATthetaselect(BAT *b, B
gdk_export BAT *BATconstant(oid hseq, int tt, const void *val, BUN cnt, role_t
role);
gdk_export gdk_return BATsubcross(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT
*sl, BAT *sr, bool max_one)
__attribute__((__warn_unused_result__));
+gdk_export gdk_return BAToutercross(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT
*sl, BAT *sr, bool max_one)
+ __attribute__((__warn_unused_result__));
gdk_export gdk_return BATleftjoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT
*sl, BAT *sr, bool nil_matches, BUN estimate)
__attribute__((__warn_unused_result__));
+gdk_export gdk_return BATmarkjoin(BAT **r1p, BAT **r2p, BAT **r3p, BAT *l, BAT
*r, BAT *sl, BAT *sr, BUN estimate)
+ __attribute__((__warn_unused_result__));
gdk_export gdk_return BATouterjoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT
*sl, BAT *sr, bool nil_matches, bool match_one, BUN estimate)
__attribute__((__warn_unused_result__));
gdk_export gdk_return BATthetajoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT
*sl, BAT *sr, int op, bool nil_matches, BUN estimate)
diff --git a/gdk/gdk_cross.c b/gdk/gdk_cross.c
--- a/gdk/gdk_cross.c
+++ b/gdk/gdk_cross.c
@@ -17,29 +17,21 @@
* The result is two bats r1 and r2 which contain the OID (head
* values) of the input bats l and r.
* If max_one is set, r can have at most one row. */
-gdk_return
-BATsubcross(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT *sl, BAT *sr, bool
max_one)
+static gdk_return
+BATcrossci(BAT **r1p, BAT **r2p, struct canditer *ci1, struct canditer *ci2)
{
BAT *bn1, *bn2 = NULL;
- struct canditer ci1, ci2;
oid *restrict p;
BUN i, j;
- canditer_init(&ci1, l, sl);
- canditer_init(&ci2, r, sr);
lng timeoffset = 0;
QryCtx *qry_ctx = MT_thread_get_qry_ctx();
if (qry_ctx != NULL) {
timeoffset = (qry_ctx->starttime && qry_ctx->querytimeout) ?
(qry_ctx->starttime + qry_ctx->querytimeout) : 0;
}
- if (max_one && ci1.ncand > 0 && ci2.ncand > 1) {
- GDKerror("more than one match");
- return GDK_FAIL;
- }
-
/* first some special cases */
- if (ci1.ncand == 0 || ci2.ncand == 0) {
+ if (ci1->ncand == 0 || ci2->ncand == 0) {
if ((bn1 = BATdense(0, 0, 0)) == NULL)
return GDK_FAIL;
if (r2p) {
@@ -52,14 +44,14 @@ BATsubcross(BAT **r1p, BAT **r2p, BAT *l
*r1p = bn1;
return GDK_SUCCEED;
}
- if (ci2.ncand == 1) {
- if ((bn1 = canditer_slice(&ci1, 0, ci1.ncand)) == NULL)
+ if (ci2->ncand == 1) {
+ if ((bn1 = canditer_slice(ci1, 0, ci1->ncand)) == NULL)
return GDK_FAIL;
if (r2p) {
- if (ci1.ncand == 1) {
- bn2 = canditer_slice(&ci2, 0, ci2.ncand);
+ if (ci1->ncand == 1) {
+ bn2 = canditer_slice(ci2, 0, ci2->ncand);
} else {
- bn2 = BATconstant(0, TYPE_oid, &ci2.seq,
ci1.ncand, TRANSIENT);
+ bn2 = BATconstant(0, TYPE_oid, &ci2->seq,
ci1->ncand, TRANSIENT);
}
if (bn2 == NULL) {
BBPreclaim(bn1);
@@ -70,12 +62,12 @@ BATsubcross(BAT **r1p, BAT **r2p, BAT *l
*r1p = bn1;
return GDK_SUCCEED;
}
- if (ci1.ncand == 1) {
- bn1 = BATconstant(0, TYPE_oid, &ci1.seq, ci2.ncand, TRANSIENT);
+ if (ci1->ncand == 1) {
+ bn1 = BATconstant(0, TYPE_oid, &ci1->seq, ci2->ncand,
TRANSIENT);
if (bn1 == NULL)
return GDK_FAIL;
if (r2p) {
- bn2 = canditer_slice(&ci2, 0, ci2.ncand);
+ bn2 = canditer_slice(ci2, 0, ci2->ncand);
if (bn2 == NULL) {
BBPreclaim(bn1);
return GDK_FAIL;
@@ -86,48 +78,48 @@ BATsubcross(BAT **r1p, BAT **r2p, BAT *l
return GDK_SUCCEED;
}
- bn1 = COLnew(0, TYPE_oid, ci1.ncand * ci2.ncand, TRANSIENT);
+ bn1 = COLnew(0, TYPE_oid, ci1->ncand * ci2->ncand, TRANSIENT);
if (r2p)
- bn2 = COLnew(0, TYPE_oid, ci1.ncand * ci2.ncand, TRANSIENT);
+ bn2 = COLnew(0, TYPE_oid, ci1->ncand * ci2->ncand, TRANSIENT);
if (!bn1 || (r2p && !bn2)) {
BBPreclaim(bn1);
if (bn2)
BBPreclaim(bn2);
return GDK_FAIL;
}
- if (ci1.ncand > 0 && ci2.ncand > 0) {
- BATsetcount(bn1, ci1.ncand * ci2.ncand);
+ if (ci1->ncand > 0 && ci2->ncand > 0) {
+ BATsetcount(bn1, ci1->ncand * ci2->ncand);
bn1->tsorted = true;
- bn1->trevsorted = ci1.ncand <= 1;
- bn1->tkey = ci2.ncand <= 1;
+ bn1->trevsorted = ci1->ncand <= 1;
+ bn1->tkey = ci2->ncand <= 1;
bn1->tnil = false;
bn1->tnonil = true;
p = (oid *) Tloc(bn1, 0);
- for (i = 0; i < ci1.ncand; i++) {
+ for (i = 0; i < ci1->ncand; i++) {
GDK_CHECK_TIMEOUT_BODY(timeoffset,
GOTO_LABEL_TIMEOUT_HANDLER(bailout));
- oid x = canditer_next(&ci1);
- for (j = 0; j < ci2.ncand; j++) {
+ oid x = canditer_next(ci1);
+ for (j = 0; j < ci2->ncand; j++) {
*p++ = x;
}
}
- BATtseqbase(bn1, ci2.ncand == 1 ? *(oid *) Tloc(bn1, 0) :
oid_nil);
+ BATtseqbase(bn1, ci2->ncand == 1 ? *(oid *) Tloc(bn1, 0) :
oid_nil);
if (bn2) {
- BATsetcount(bn2, ci1.ncand * ci2.ncand);
- bn2->tsorted = ci1.ncand <= 1 || ci2.ncand <= 1;
- bn2->trevsorted = ci2.ncand <= 1;
- bn2->tkey = ci1.ncand <= 1;
+ BATsetcount(bn2, ci1->ncand * ci2->ncand);
+ bn2->tsorted = ci1->ncand <= 1 || ci2->ncand <= 1;
+ bn2->trevsorted = ci2->ncand <= 1;
+ bn2->tkey = ci1->ncand <= 1;
bn2->tnil = false;
bn2->tnonil = true;
p = (oid *) Tloc(bn2, 0);
- for (i = 0; i < ci1.ncand; i++) {
+ for (i = 0; i < ci1->ncand; i++) {
GDK_CHECK_TIMEOUT_BODY(timeoffset,
GOTO_LABEL_TIMEOUT_HANDLER(bailout));
- for (j = 0; j < ci2.ncand; j++) {
- *p++ = canditer_next(&ci2);
+ for (j = 0; j < ci2->ncand; j++) {
+ *p++ = canditer_next(ci2);
}
- canditer_reset(&ci2);
+ canditer_reset(ci2);
}
- BATtseqbase(bn2, ci1.ncand == 1 ? *(oid *) Tloc(bn2, 0)
: oid_nil);
+ BATtseqbase(bn2, ci1->ncand == 1 ? *(oid *) Tloc(bn2,
0) : oid_nil);
_______________________________________________
checkin-list mailing list -- checkin-list@monetdb.org
To unsubscribe send an email to checkin-list-le...@monetdb.org
