Update of /cvsroot/monetdb/sql/src/backends/monet5
In directory 23jxhf1.ch3.sourceforge.com:/tmp/cvs-serv15984/src/backends/monet5
Modified Files:
sql.mx sql_gencode.mx
Log Message:
added support for sql/xml namespaces
fixed bug in algebra version for handling exist/not exists in exists/not exists
with correlation
added optimizer steps for algebra version to rewrite semi/anti joins followed
by joins on the same (referenced) relation into a single semijoin
push selects through semi/anti joins
approved tests
U sql_gencode.mx
Index: sql_gencode.mx
===================================================================
RCS file: /cvsroot/monetdb/sql/src/backends/monet5/sql_gencode.mx,v
retrieving revision 1.289
retrieving revision 1.290
diff -u -d -r1.289 -r1.290
--- sql_gencode.mx 10 Oct 2008 08:54:31 -0000 1.289
+++ sql_gencode.mx 28 Oct 2008 20:05:04 -0000 1.290
@@ -1267,7 +1267,14 @@
int l = _dumpstmt(sql, mb, s->op1.stval);
backend_create_func(sql, s->op4.funcval->func);
- if (s->op1.stval->nrcols) {
+ if (s->op1.stval->nrcols && strcmp(fimp, "not_uniques")
== 0) {
+ int rtype = s->op4.funcval->res.type->localtype;
+
+ q = newStmt(mb, mod, fimp);
+ setVarType(mb,getArg(q,0),
newBatType(TYPE_oid,rtype));
+ setVarUDFtype(mb,getArg(q,0));
+ q = pushArgument(mb, q, l);
+ } else if (s->op1.stval->nrcols) {
int rtype = s->op4.funcval->res.type->localtype;
q = newStmt(mb, "mal","multiplex");
U sql.mx
Index: sql.mx
===================================================================
RCS file: /cvsroot/monetdb/sql/src/backends/monet5/sql.mx,v
retrieving revision 1.289
retrieving revision 1.290
diff -u -d -r1.289 -r1.290
--- sql.mx 10 Oct 2008 08:54:23 -0000 1.289
+++ sql.mx 28 Oct 2008 20:05:04 -0000 1.290
@@ -230,8 +230,12 @@
command not_unique( b:bat[:oid,:oid]) :bit
address not_unique
comment "check if the tail sorted bat b doesn't have unique tail values" ;
-command not_unique( b:bat[:oid,:oid]) :bit
-address not_unique ;
+
+command not_uniques( b:bat[:oid,:oid]) :bat[:oid,:oid]
+address not_unique_oids
+comment "return not unique oids" ;
+command not_uniques( b:bat[:oid,:wrd]) :bat[:oid,:oid]
+address not_unique_oids ;
function sql_environment():bat[:str,:bat];
b := bat.new(:str,:bat,2);
@@ -758,6 +762,7 @@
sql5_export str mvc_restart_seq(Client cntxt, MalBlkPtr mb, MalStkPtr stk,
InstrPtr pci);
sql5_export str zero_or_one(ptr ret, int *bid);
sql5_export str not_unique(bit *ret, int *bid);
+sql5_export str not_unique_oids(bat *ret, bat *bid);
sql5_export str month_interval_str( int *ret, str *s, int *ek, int *sk );
sql5_export str second_interval_str( lng *res, str *s, int *ek, int *sk );
sql5_export str dump_cache(Client cntxt, MalBlkPtr mb, MalStkPtr stk, InstrPtr
pci);
@@ -2096,6 +2101,96 @@
return MAL_SUCCEED;
}
+/* later we could optimize this to start from current BUN
+ And only search the from the first if second is not found.
+ */
+inline int
+HASHfndTwice(BAT *b, ptr v)
+{
+ BATiter bi = bat_iterator(b);
+ BUN i = BUN_NONE;
+ int first = 1;
+
+ HASHloop( bi, b->H->hash, i, v) {
+ if (!first)
+ return 1;
+ first = 0;
+ }
+ return 0;
+}
+
+str
+not_unique_oids(bat *ret, bat* bid)
+{
+ BAT *b, *bn = NULL;
+
+ if ((b = BATdescriptor(*bid)) == NULL) {
+ throw(SQL, "not_uniques", "Cannot access descriptor");
+ }
+ if (b->ttype != TYPE_oid && b->ttype != TYPE_wrd) {
+ throw(SQL, "not_uniques", "Wrong types");
+ }
+
+ assert(b->htype == TYPE_oid);
+ if (BATtkey(b) || BATtdense(b) || BATcount(b) <= 1) {
+ bn = BATnew(TYPE_oid, TYPE_oid, 0);
+ } else if (b->tsorted&1) { /* ugh handle both wrd and oid types */
+ oid c = *(oid*)Tloc(b, BUNfirst(b)), *rf, *rh, *rt;
+ oid *h = (oid*)Hloc(b,0), *vp, *ve;
+ int first = 1;
+
+ bn = BATnew(TYPE_oid, TYPE_oid, BATcount(b));
+ vp = (oid*)Tloc(b, BUNfirst(b));
+ ve = vp + BATcount(b);
+ rf = rh = (oid*)Hloc(bn, BUNfirst(bn));
+ rt = (oid*)Tloc(bn, BUNfirst(bn));
+ *rh++ = *h++;
+ *rt++ = *vp;
+ for(vp++; vp < ve; vp++, h++) {
+ oid v = *vp;
+ if (v == c) {
+ first = 0;
+ *rh++ = *h;
+ *rt++ = v;
+ } else if (!first) {
+ first = 1;
+ *rh++ = *h;
+ *rt++ = v;
+ } else {
+ *rh = *h;
+ *rt = v;
+ }
+ c = v;
+ }
+ if (first)
+ rh--;
+ BATsetcount(bn, rh - rf);
+ } else {
+ oid *rf, *rh, *rt;
+ oid *h = (oid*)Hloc(b,0), *vp, *ve;
+ BAT *bm = BATmirror(b);
+
+ if (BATprepareHash(bm))
+ throw(SQL, "not_uniques", "hash creation failed");
+ bn = BATnew(TYPE_oid, TYPE_oid, BATcount(b));
+ vp = (oid*)Tloc(b, BUNfirst(b));
+ ve = vp + BATcount(b);
+ rf = rh = (oid*)Hloc(bn, BUNfirst(bn));
+ rt = (oid*)Tloc(bn, BUNfirst(bn));
+ for(; vp < ve; vp++, h++) {
+ /* try to find value twice */
+ if (HASHfndTwice(bm, vp)) {
+ *rh++ = *h;
+ *rt++ = *vp;
+ }
+ }
+ BATsetcount(bn, rh - rf);
+ }
+ BBPunfix(b->batCacheid);
+ BBPkeepref(*ret = bn->batCacheid);
+ return MAL_SUCCEED;
+}
+
static str
voidbathash(bat *res, BAT *b )
{
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Monetdb-sql-checkins mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-sql-checkins
