Changeset: 29486adb8cba for MonetDB
URL: https://dev.monetdb.org/hg/MonetDB/rev/29486adb8cba
Modified Files:
sql/server/sql_env.c
sql/server/sql_mvc.c
sql/server/sql_mvc.h
sql/test/BugDay_2005-10-06_2.9.3/Tests/CrashMe_SQL_server_crash-2.SF-921673.test
sql/test/Dependencies/Tests/dependency_owner_schema_3.test
sql/test/bugs/Tests/innerjoin_multiple-bug-sf-943661.test
Branch: ordered-set-aggregates
Log Message:
merged with default
diffs (truncated from 27932 to 300 lines):
diff --git a/clients/mapiclient/mhelp.c b/clients/mapiclient/mhelp.c
--- a/clients/mapiclient/mhelp.c
+++ b/clients/mapiclient/mhelp.c
@@ -68,7 +68,7 @@ SQLhelp sqlhelp1[] = {
{"ALTER TABLE",
"",
"ALTER TABLE [ IF EXISTS ] qname ADD [ COLUMN ] column_def\n"
- "ALTER TABLE [ IF EXISTS ] qname ADD table_constraint\n"
+ "ALTER TABLE [ IF EXISTS ] qname ADD [ CONSTRAINT ident ]
table_constraint\n"
"ALTER TABLE [ IF EXISTS ] qname ALTER [ COLUMN ] ident SET DEFAULT
value\n"
"ALTER TABLE [ IF EXISTS ] qname ALTER [ COLUMN ] ident SET [NOT]
NULL\n"
"ALTER TABLE [ IF EXISTS ] qname ALTER [ COLUMN ] ident DROP DEFAULT\n"
@@ -910,9 +910,10 @@ SQLhelp sqlhelp2[] = {
NULL,},
{"table_constraint",
NULL,
- "[ CONSTRAINT ident ] { PRIMARY KEY column_list | UNIQUE column_list
|\n"
+ "[ CONSTRAINT ident ] { CHECK '(' search_condition ')' |\n"
+ " PRIMARY KEY column_list | UNIQUE column_list |\n"
" FOREIGN KEY column_list REFERENCES qname [ column_list ] [
match_options ] [ reference_action ] }",
- "column_list,match_options,reference_action",
+ "column_list,search_condition,match_options,reference_action",
"See also
https://www.monetdb.org/documentation/user-guide/sql-manual/data-definition/table-elements/"},
{"table_element",
NULL,
diff --git a/clients/mapilib/mapi.c b/clients/mapilib/mapi.c
--- a/clients/mapilib/mapi.c
+++ b/clients/mapilib/mapi.c
@@ -1217,7 +1217,7 @@ mapi_log_header(Mapi mid, const char *fu
if (firstcall == 0)
firstcall = now;
double seconds = (double)(now - firstcall) / 1e6;
- mnstr_printf(mid->tracelog, "\342\226\266 [%u] t=%.3fs %s%s %s(), line
%ld\n", mid->index, seconds, mark1, mark2, funcname, line); /* U+25B6:
right-pointing triangle */
+ mnstr_printf(mid->tracelog, "\n** [%u] t=%.3fs %s%s %s(), line %ld\n",
mid->index, seconds, mark1, mark2, funcname, line);
}
void
diff --git a/gdk/gdk_join.c b/gdk/gdk_join.c
--- a/gdk/gdk_join.c
+++ b/gdk/gdk_join.c
@@ -4212,7 +4212,7 @@ leftjoin(BAT **r1p, BAT **r2p, BAT **r3p
/* maybe do a hash join on the swapped operands; if we
* do, we need to sort the output, so we take that into
* account as well */
- bool lhash, plhash, lcand;
+ bool lhash, plhash, lcand, rkey = r->tkey;
double lcost;
lcost = joincost(l, rci.ncand, &lci, &lhash, &plhash, &lcand);
@@ -4220,7 +4220,7 @@ leftjoin(BAT **r1p, BAT **r2p, BAT **r3p
rc = GDK_FAIL;
goto doreturn;
}
- if (semi)
+ if (semi && !rkey)
lcost += rci.ncand; /* cost of BATunique(r) */
/* add cost of sorting; obviously we don't know the
* size, so we guess that the size of the output is
@@ -4229,7 +4229,7 @@ leftjoin(BAT **r1p, BAT **r2p, BAT **r3p
if (lcost < rcost) {
BAT *tmp = sr;
BAT *r1, *r2;
- if (semi) {
+ if (semi && !rkey) {
sr = BATunique(r, sr);
if (sr == NULL) {
rc = GDK_FAIL;
@@ -4240,7 +4240,7 @@ leftjoin(BAT **r1p, BAT **r2p, BAT **r3p
rc = hashjoin(&r2, &r1, NULL, r, l, &rci, &lci,
nil_matches,
false, false, false, false, false, false,
estimate,
t0, true, lhash, plhash, lcand, func);
- if (semi)
+ if (semi && !rkey)
BBPunfix(sr->batCacheid);
if (rc != GDK_SUCCEED)
goto doreturn;
diff --git a/gdk/gdk_select.c b/gdk/gdk_select.c
--- a/gdk/gdk_select.c
+++ b/gdk/gdk_select.c
@@ -1986,30 +1986,27 @@ BATselect(BAT *b, BAT *s, const void *tl
MT_lock_unset(&pb->batIdxLock);
}
if (oidxh) {
- /* Is query selective enough to use the ordered index ?
*/
- /* TODO: Test if this heuristic works in practice */
- /*if ((ORDERfnd(b, th) - ORDERfnd(b, tl)) < ((BUN)1000
< bi.count/1000 ? (BUN)1000: bi.count/1000))*/
- if ((ORDERfnd(b, oidxh, th) - ORDERfnd(b, oidxh, tl)) <
bi.count/3) {
- if (view) {
- bat_iterator_end(&bi);
- bi = bat_iterator(b);
- poidx = true; /* using parent oidx */
- vwo = (lng) (view->tbaseoff -
bi.baseoff);
- vwl = b->hseqbase + (oid) vwo + ci.seq
- view->hseqbase;
- vwh = vwl + canditer_last(&ci) - ci.seq;
- vwo = (lng) view->hseqbase - (lng)
b->hseqbase - vwo;
- TRC_DEBUG(ALGO, "Switch from "
ALGOBATFMT " to " ALGOBATFMT " " OIDFMT "-" OIDFMT " hseq " LLFMT "\n",
ALGOBATPAR(view), ALGOBATPAR(b), vwl, vwh, vwo);
- } else {
- vwl = ci.seq;
- vwh = canditer_last(&ci);
- }
+ /* Is query selective enough to use the ordered
+ * index? Finding the boundaries is 2*log(n)
+ * where n is the size of the bat, sorting is
+ * N*log(N) where N is the number of results.
+ * If the sum is less than n (cost of scan),
+ * it's cheaper. However, to find out how large
+ * N is, we'd have to do the two boundary
+ * searches. If we do that, we might as well do
+ * it all. */
+ if (view) {
+ bat_iterator_end(&bi);
+ bi = bat_iterator(b);
+ poidx = true; /* using parent oidx */
+ vwo = (lng) (view->tbaseoff - bi.baseoff);
+ vwl = b->hseqbase + (oid) vwo + ci.seq -
view->hseqbase;
+ vwh = vwl + canditer_last(&ci) - ci.seq;
+ vwo = (lng) view->hseqbase - (lng) b->hseqbase
- vwo;
+ TRC_DEBUG(ALGO, "Switch from " ALGOBATFMT " to
" ALGOBATFMT " " OIDFMT "-" OIDFMT " hseq " LLFMT "\n", ALGOBATPAR(view),
ALGOBATPAR(b), vwl, vwh, vwo);
} else {
- if (view) {
- b = view;
- view = NULL;
- }
- HEAPdecref(oidxh, false);
- oidxh = NULL;
+ vwl = ci.seq;
+ vwh = canditer_last(&ci);
}
}
}
diff --git a/geom/sql/conformance/Tests/T1.test
b/geom/sql/conformance/Tests/T1.test
--- a/geom/sql/conformance/Tests/T1.test
+++ b/geom/sql/conformance/Tests/T1.test
@@ -1,4 +1,4 @@
-query T rowsort
+query T nosort
SELECT f_table_name FROM geometry_columns ORDER BY f_table_name
----
bridges
diff --git a/sql/ChangeLog b/sql/ChangeLog
--- a/sql/ChangeLog
+++ b/sql/ChangeLog
@@ -1,3 +1,7 @@
# ChangeLog file for sql
# This file is updated with Maddlog
+* Tue Sep 10 2024 Lucas Pereira <lucas.pere...@monetdbsolutions.com>
+- Introduce division_min_scale SQL environment variable for specifying
+ minimum scale of the division result. The default value is 3.
+
diff --git a/sql/backends/monet5/Tests/rapi05.test
b/sql/backends/monet5/Tests/rapi05.test
--- a/sql/backends/monet5/Tests/rapi05.test
+++ b/sql/backends/monet5/Tests/rapi05.test
@@ -11,7 +11,7 @@ statement ok
CREATE FUNCTION kmeans(data float, ncluster integer) returns integer
language R {kmeans(data,ncluster)$cluster}
-query I rowsort
+query I nosort
SELECT cluster FROM (SELECT MIN(x) AS minx, MAX(x) AS maxx, kmeans(x,5) AS
cluster FROM xdata
GROUP BY cluster) as cdata ORDER BY cluster
----
diff --git a/sql/backends/monet5/generator/Tests/joins00.test
b/sql/backends/monet5/generator/Tests/joins00.test
--- a/sql/backends/monet5/generator/Tests/joins00.test
+++ b/sql/backends/monet5/generator/Tests/joins00.test
@@ -98,14 +98,14 @@ create table tmp(i tinyint)
statement ok
insert into tmp values(3),(4),(5)
-query I rowsort
+query I nosort
select * from tmp order by i
----
3
4
5
-query II rowsort
+query II nosort
select * from generate_series(9,0,-2) X, tmp Y where X.value = Y.i order by
X.value, Y.i
----
3
@@ -113,7 +113,7 @@ 3
5
5
-query II rowsort
+query II nosort
select * from generate_series(9,0,-2) X, tmp Y where Y.i = X.value order by
X.value, Y.i
----
3
@@ -121,13 +121,13 @@ 3
5
5
-query II rowsort
+query II nosort
select * from generate_series(9,0,-2) X, tmp Y where X.value = Y.i and value
<5 order by X.value, Y.i
----
3
3
-query II rowsort
+query II nosort
select * from generate_series(9,0,-2) as X, tmp Y where X.value = Y.i and
value <7 and value >3 order by X.value, Y.i
----
5
diff --git a/sql/backends/monet5/generator/Tests/joins01.test
b/sql/backends/monet5/generator/Tests/joins01.test
--- a/sql/backends/monet5/generator/Tests/joins01.test
+++ b/sql/backends/monet5/generator/Tests/joins01.test
@@ -7,14 +7,14 @@ insert into tmp values
(timestamp '2008-03-01 10:00'),
(timestamp '2008-03-01 20:00')
-query T rowsort
+query T nosort
select * from tmp order by i
----
2008-03-01 00:00:00
2008-03-01 10:00:00
2008-03-01 20:00:00
-query TT rowsort
+query TT nosort
select * from generate_series(timestamp '2008-03-01 00:00',timestamp
'2008-03-04 12:00',cast( '10' as interval hour)) X, tmp Y where X.value = Y.i
order by X.value, Y.i
----
2008-03-01 00:00:00
@@ -24,7 +24,7 @@ 2008-03-01 10:00:00
2008-03-01 20:00:00
2008-03-01 20:00:00
-query TT rowsort
+query TT nosort
select * from generate_series(timestamp '2008-03-01 00:00',timestamp
'2008-03-04 12:00',cast( '10' as interval hour)) X, tmp Y where Y.i = X.value
order by X.value, Y.i
----
2008-03-01 00:00:00
@@ -34,7 +34,7 @@ 2008-03-01 10:00:00
2008-03-01 20:00:00
2008-03-01 20:00:00
-query TT rowsort
+query TT nosort
select * from generate_series(timestamp '2008-03-01 00:00',timestamp
'2008-03-04 12:00',cast( '10' as interval hour)) X, tmp Y where X.value = Y.i
and value < timestamp '2008-03-01 20:00' order by X.value, Y.i
----
2008-03-01 00:00:00
@@ -42,7 +42,7 @@ 2008-03-01 00:00:00
2008-03-01 10:00:00
2008-03-01 10:00:00
-query TT rowsort
+query TT nosort
select * from generate_series(timestamp '2008-03-01 00:00',timestamp
'2008-03-04 12:00',cast( '10' as interval hour)) X, tmp Y where Y.i = X.value
and value < timestamp '2008-03-01 20:00' order by X.value, Y.i
----
2008-03-01 00:00:00
@@ -50,7 +50,7 @@ 2008-03-01 00:00:00
2008-03-01 10:00:00
2008-03-01 10:00:00
-query TT rowsort
+query TT nosort
select * from generate_series(timestamp '2008-03-01 00:00',timestamp
'2008-03-04 12:00',cast( '10' as interval hour)) as X, tmp Y where X.value =
Y.i and value < timestamp '2008-03-01 20:00' and value > timestamp '200-03-01
00:00' order by X.value, Y.i
----
2008-03-01 00:00:00
@@ -58,7 +58,7 @@ 2008-03-01 00:00:00
2008-03-01 10:00:00
2008-03-01 10:00:00
-query TT rowsort
+query TT nosort
select * from generate_series(timestamp '2008-03-01 00:00',timestamp
'2008-03-04 12:00',cast( '10' as interval hour)) as X, tmp Y where X.value =
Y.i and i < timestamp '2008-03-01 20:00' and i > timestamp '200-03-01 00:00'
order by X.value, Y.i
----
2008-03-01 00:00:00
@@ -66,7 +66,7 @@ 2008-03-01 00:00:00
2008-03-01 10:00:00
2008-03-01 10:00:00
-query T rowsort
+query T nosort
select * from generate_series(timestamp '2008-03-04 18:00',timestamp
'2008-03-01 00:00',cast( '-10' as interval hour)) X order by X.value
----
2008-03-01 10:00:00
@@ -79,7 +79,7 @@ 2008-03-03 22:00:00
2008-03-04 08:00:00
2008-03-04 18:00:00
-query TT rowsort
+query TT nosort
select * from generate_series(timestamp '2008-03-04 18:00',timestamp
'2008-03-01 00:00',cast( '-10' as interval hour)) X, tmp Y where X.value = Y.i
order by X.value, Y.i
----
2008-03-01 10:00:00
@@ -87,7 +87,7 @@ 2008-03-01 10:00:00
2008-03-01 20:00:00
2008-03-01 20:00:00
-query TT rowsort
+query TT nosort
select * from generate_series(timestamp '2008-03-04 18:00',timestamp
'2008-03-01 00:00',cast( '-10' as interval hour)) X, tmp Y where Y.i = X.value
order by X.value, Y.i
----
2008-03-01 10:00:00
@@ -95,26 +95,26 @@ 2008-03-01 10:00:00
_______________________________________________
checkin-list mailing list -- checkin-list@monetdb.org
To unsubscribe send an email to checkin-list-le...@monetdb.org
