Changeset: 3dd6bf4ad4f4 for MonetDB
URL: https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=3dd6bf4ad4f4
Modified Files:
clients/Tests/exports.stable.out
gdk/gdk.h
gdk/gdk_firstn.c
gdk/gdk_join.c
monetdb5/modules/kernel/algebra.c
sql/backends/monet5/sql.c
Branch: default
Log Message:
New function BATintersect.
This is equivalent to BATsemijoin with a single output and similar to BATdiff.
diffs (174 lines):
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
@@ -139,6 +139,7 @@ BUN BATgrows(BAT *b);
gdk_return BAThash(BAT *b, BUN masksize);
void BAThseqbase(BAT *b, oid o);
gdk_return BATimprints(BAT *b);
+BAT *BATintersect(BAT *l, BAT *r, BAT *sl, BAT *sr, int nil_matches, BUN
estimate);
BAT *BATintersectcand(BAT *a, BAT *b);
gdk_return BATjoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT *sl, BAT *sr, int
nil_matches, BUN estimate) __attribute__((__warn_unused_result__));
gdk_return BATkey(BAT *b, bool onoff);
diff --git a/gdk/gdk.h b/gdk/gdk.h
--- a/gdk/gdk.h
+++ b/gdk/gdk.h
@@ -2746,6 +2746,7 @@ gdk_export gdk_return BATthetajoin(BAT *
__attribute__ ((__warn_unused_result__));
gdk_export gdk_return BATsemijoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT
*sl, BAT *sr, int nil_matches, BUN estimate)
__attribute__ ((__warn_unused_result__));
+gdk_export BAT *BATintersect(BAT *l, BAT *r, BAT *sl, BAT *sr, int
nil_matches, BUN estimate);
gdk_export BAT *BATdiff(BAT *l, BAT *r, BAT *sl, BAT *sr, int nil_matches, BUN
estimate);
gdk_export gdk_return BATjoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT *sl,
BAT *sr, int nil_matches, BUN estimate)
__attribute__ ((__warn_unused_result__));
diff --git a/gdk/gdk_firstn.c b/gdk/gdk_firstn.c
--- a/gdk/gdk_firstn.c
+++ b/gdk/gdk_firstn.c
@@ -653,9 +653,9 @@ BATfirstn_grouped(BAT **topn, BAT **gids
bn1 = bn;
BBPunfix(s->batCacheid);
- rc = BATsemijoin(&bn, NULL, b, b, su, bn1, 1, BUN_NONE);
+ bn = BATintersect(b, b, su, bn1, 1, BUN_NONE);
BBPunfix(bn1->batCacheid);
- if (rc != GDK_SUCCEED)
+ if (bn == NULL)
return GDK_FAIL;
} else {
BATiter bi = bat_iterator(b);
@@ -725,9 +725,9 @@ BATfirstn_grouped_with_groups(BAT **topn
BBPunfix(bn2->batCacheid);
return GDK_FAIL;
}
- rc = BATsemijoin(&bn4, NULL, s, bn2, NULL, NULL, 0, BUN_NONE);
+ bn4 = BATintersect(s, bn2, NULL, NULL, 0, BUN_NONE);
BBPunfix(bn2->batCacheid);
- if (rc != GDK_SUCCEED) {
+ if (bn4 == NULL) {
BBPunfix(bn1->batCacheid);
return GDK_FAIL;
}
@@ -801,7 +801,7 @@ BATfirstn_grouped_with_groups(BAT **topn
if (gids) {
BAT *bn1, *bn2, *bn3, *bn4, *bn5, *bn6, *bn7, *bn8;
- if (BATsemijoin(&bn1, NULL, s, bn, NULL, NULL, 0, BUN_NONE) !=
GDK_SUCCEED) {
+ if ((bn1 = BATintersect(s, bn, NULL, NULL, 0, BUN_NONE)) ==
NULL) {
BBPunfix(bn->batCacheid);
return GDK_FAIL;
}
diff --git a/gdk/gdk_join.c b/gdk/gdk_join.c
--- a/gdk/gdk_join.c
+++ b/gdk/gdk_join.c
@@ -60,12 +60,14 @@
* right values; two extra Boolean parameters, li and hi,
* indicate whether equal values match
*
- * In addition to these functions, there is one more functions that is
- * closely related:
+ * In addition to these functions, there are two more functions that
+ * are closely related:
+ * BATintersect
+ * intersection: return a candidate list with OIDs of tuples in
+ * the left input whose value occurs in the right input
* BATdiff
- * difference: return a candidate list compatible list of OIDs of
- * tuples in the left input whose value does not occur in the
- * right input
+ * difference: return a candidate list with OIDs of tuples in the
+ * left input whose value does not occur in the right input
*/
/* Perform a bunch of sanity checks on the inputs to a join. */
@@ -3864,9 +3866,10 @@ BATouterjoin(BAT **r1p, BAT **r2p, BAT *
GDKdebug & ALGOMASK ? GDKusec() : 0);
}
-/* Perform a semi-join over l and r. Returns two new, aligned, bats
+/* Perform a semi-join over l and r. Returns one or two new, bats
* with the oids of matching tuples. The result is in the same order
- * as l (i.e. r1 is sorted). */
+ * as l (i.e. r1 is sorted). If a single bat is returned, it is a
+ * candidate list. */
gdk_return
BATsemijoin(BAT **r1p, BAT **r2p, BAT *l, BAT *r, BAT *sl, BAT *sr, int
nil_matches, BUN estimate)
{
@@ -3875,10 +3878,23 @@ BATsemijoin(BAT **r1p, BAT **r2p, BAT *l
GDKdebug & ALGOMASK ? GDKusec() : 0);
}
+/* Return a candidate list with the list of rows in l whose value also
+ * occurs in r. This is just the left output of a semi-join. */
+BAT *
+BATintersect(BAT *l, BAT *r, BAT *sl, BAT *sr, int nil_matches, BUN estimate)
+{
+ BAT *bn;
+
+ if (leftjoin(&bn, NULL, l, r, sl, sr, nil_matches,
+ false, true, false, estimate, "BATintersect",
+ GDKdebug & ALGOMASK ? GDKusec() : 0) == GDK_SUCCEED)
+ return virtualize(bn);
+ return NULL;
+}
+
/* Return the difference of l and r. The result is a BAT with the
* oids of those values in l that do not occur in r. This is what you
- * might call an anti-semi-join. The result can be used as a
- * candidate list. */
+ * might call an anti-semi-join. The result is a candidate list. */
BAT *
BATdiff(BAT *l, BAT *r, BAT *sl, BAT *sr, int nil_matches, BUN estimate)
{
@@ -3887,7 +3903,7 @@ BATdiff(BAT *l, BAT *r, BAT *sl, BAT *sr
if (leftjoin(&bn, NULL, l, r, sl, sr, nil_matches,
false, false, true, estimate, "BATdiff",
GDKdebug & ALGOMASK ? GDKusec() : 0) == GDK_SUCCEED)
- return bn;
+ return virtualize(bn);
return NULL;
}
diff --git a/monetdb5/modules/kernel/algebra.c
b/monetdb5/modules/kernel/algebra.c
--- a/monetdb5/modules/kernel/algebra.c
+++ b/monetdb5/modules/kernel/algebra.c
@@ -508,7 +508,7 @@ ALGintersect(bat *r1, const bat *lid, co
{
return do_join(r1, NULL, lid, rid, NULL, slid, srid, 0, NULL, NULL, 0,
0,
nil_matches, estimate,
- BATsemijoin, NULL, NULL, NULL, NULL,
"algebra.intersect");
+ NULL, NULL, NULL, NULL, BATintersect,
"algebra.intersect");
}
/* algebra.firstn(b:bat[:any],
diff --git a/sql/backends/monet5/sql.c b/sql/backends/monet5/sql.c
--- a/sql/backends/monet5/sql.c
+++ b/sql/backends/monet5/sql.c
@@ -1537,9 +1537,9 @@ DELTAsub(bat *result, const bat *col, co
BBPunfix(u->batCacheid);
throw(MAL, "sql.delta", SQLSTATE(HY002)
RUNTIME_OBJECT_MISSING);
}
- ret = BATsemijoin(&cminu, NULL, u, c_ids, NULL, NULL,
0, BUN_NONE);
+ cminu = BATintersect(u, c_ids, NULL, NULL, 0, BUN_NONE);
BBPunfix(c_ids->batCacheid);
- if (ret != GDK_SUCCEED) {
+ if (cminu == NULL) {
BBPunfix(c->batCacheid);
BBPunfix(u->batCacheid);
throw(MAL, "sql.delta", SQLSTATE(HY001)
MAL_MALLOC_FAIL);
@@ -1704,7 +1704,7 @@ DELTAproject(bat *result, const bat *sub
/* create subsets of u_id and u_val where the tail
* values of u_id are also in s, and where those tail
* values occur as head value in res */
- if (BATsemijoin(&o, NULL, u_id, s, NULL, NULL, 0, BUN_NONE) !=
GDK_SUCCEED) {
+ if ((o = BATintersect(u_id, s, NULL, NULL, 0, BUN_NONE)) ==
NULL) {
BBPunfix(s->batCacheid);
BBPunfix(res->batCacheid);
BBPunfix(u_id->batCacheid);
@@ -1720,7 +1720,7 @@ DELTAproject(bat *result, const bat *sub
if (nu_id == NULL ||
nu_val == NULL ||
tres == NULL ||
- BATsemijoin(&o, NULL, nu_id, tres, NULL, NULL, 0, BUN_NONE)
!= GDK_SUCCEED) {
+ (o = BATintersect(nu_id, tres, NULL, NULL, 0, BUN_NONE)) ==
NULL) {
BBPunfix(s->batCacheid);
BBPunfix(res->batCacheid);
BBPreclaim(nu_id);
_______________________________________________
checkin-list mailing list
checkin-list@monetdb.org
https://www.monetdb.org/mailman/listinfo/checkin-list
