Changeset: 0c3d286ac863 for MonetDB
URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=0c3d286ac863
Modified Files:
monetdb5/extras/rdf/hashmap/hashmap.c
monetdb5/extras/rdf/hashmap/hashmap.h
monetdb5/extras/rdf/rdfschema.c
Branch: rdf
Log Message:
Modify the function for storing predicates into a hash (putPtoHash).
diffs (truncated from 314 to 300 lines):
diff --git a/monetdb5/extras/rdf/hashmap/hashmap.c
b/monetdb5/extras/rdf/hashmap/hashmap.c
--- a/monetdb5/extras/rdf/hashmap/hashmap.c
+++ b/monetdb5/extras/rdf/hashmap/hashmap.c
@@ -1,6 +1,5 @@
-/*
- * Generic map implementation.
- */
+/* HashMap for the characteristic sets (CSs') in RDF */
+
#include <hashmap.h>
#include <stdlib.h>
@@ -62,6 +61,37 @@ static char intsetcmp(int* key1, int* ke
return 0;
}
+/* Return 1 if sorted arr2[] is a subset of sorted arr1[]
+ * arr1 has m members, arr2 has n members
+ * */
+
+static int isSubset(int* arr1, int* arr2, int m, int n)
+{
+ int i = 0, j = 0;
+
+ if(m < n)
+ return 0;
+
+ while( i < n && j < m )
+ {
+ if( arr1[j] < arr2[i] )
+ j++;
+ else if( arr1[j] == arr2[i] )
+ {
+ j++;
+ i++;
+ }
+ else if( arr1[j] > arr2[i] )
+ return 0;
+ }
+
+ if( i < n )
+ return 0;
+ else
+ return 1;
+}
+
+
/*
* Return the integer of the location in data
* to store the point to the item, or MAP_FULL.
@@ -163,6 +193,39 @@ int hashmap_put(map_t in, int* key, int
return MAP_OK;
}
+
+/*
+ * Add a pointer to the hashmap with some key
+ * This function is ONLY used for the case of
+ * the predicate in RDF triple
+ */
+int hashmap_put_forP(map_t in, int* key, int num, any_t value, int support){
+ int index;
+ hashmap_map* m;
+
+ /* Cast the hashmap */
+ m = (hashmap_map *) in;
+
+ /* Find a place to put our value */
+ index = hashmap_hash(in, key, num);
+ while(index == MAP_FULL){
+ if (hashmap_rehash(in) == MAP_OMEM) {
+ return MAP_OMEM;
+ }
+ index = hashmap_hash(in, key, num);
+ }
+
+ /* Set the data */
+ m->data[index].data = value;
+ m->data[index].key = key;
+ m->data[index].num = num;
+ m->data[index].in_use = 1;
+ m->data[index].freq = support;
+ m->size++;
+
+ return MAP_OK;
+}
+
/*
* Get your pointer out of the hashmap with a key
*/
@@ -180,7 +243,7 @@ int hashmap_get(map_t in, int* key, int
/* Linear probing, if necessary */
for(i = 0; i<MAX_CHAIN_LENGTH; i++){
- int in_use = m->data[curr].in_use;
+ char in_use = m->data[curr].in_use;
if (in_use == 1){
if ((m->data[curr].num == num) &&
(intsetcmp(m->data[curr].key,key,num)==0)){
*arg = (m->data[curr].data);
@@ -201,6 +264,42 @@ int hashmap_get(map_t in, int* key, int
}
+/*
+ * Get your pointer out of the hashmap with a key
+ * This function is ONLY used for the case of
+ * the predicate in RDF triple
+ */
+int hashmap_get_forP(map_t in, int* key, any_t *arg){
+ int curr;
+ int i;
+ hashmap_map* m;
+
+ /* Cast the hashmap */
+ m = (hashmap_map *) in;
+
+ /* Find data location */
+ curr = hashmap_hash_int(m, key, 1);
+
+ /* Linear probing, if necessary */
+ for(i = 0; i<MAX_CHAIN_LENGTH; i++){
+
+ int in_use = m->data[curr].in_use;
+ if (in_use == 1){
+ if (*m->data[curr].key == *key){
+ *arg = (m->data[curr].data);
+ return MAP_OK;
+ }
+ }
+ curr = (curr + 1) % m->table_size;
+ }
+
+ *arg = NULL;
+
+ /* Not found */
+ return MAP_MISSING;
+}
+
+
/*
* Iterate the function parameter over each element in the hashmap. The
@@ -367,7 +466,7 @@ int hashmap_remove(map_t in, int* key, i
/* Linear probing, if necessary */
for(i = 0; i<MAX_CHAIN_LENGTH; i++){
- int in_use = m->data[curr].in_use;
+ char in_use = m->data[curr].in_use;
if (in_use == 1){
if ((m->data[curr].num == num) &&
(intsetcmp(m->data[curr].key,key,num)==0)){
/* Blank out the fields */
diff --git a/monetdb5/extras/rdf/hashmap/hashmap.h
b/monetdb5/extras/rdf/hashmap/hashmap.h
--- a/monetdb5/extras/rdf/hashmap/hashmap.h
+++ b/monetdb5/extras/rdf/hashmap/hashmap.h
@@ -1,13 +1,9 @@
-/*
- * Generic hashmap manipulation functions
- *
- * Originally by Elliot C Back -
http://elliottback.com/wp/hashmap-implementation-in-c/
- *
- * Modified by Pete Warden to fix a serious performance problem, support
strings as keys
- * and removed thread synchronization - http://petewarden.typepad.com
- *
- * Modified by Minh-Duc Pham to support key as a set of integer values
- * Added functions for collecting data statistic
+/* HashMap for the characteristic sets (CSs') in RDF */
+
+/* Originally by Elliot C Back -
http://elliottback.com/wp/hashmap-implementation-in-c/
+ * Modified by Minh-Duc Pham to support key as a set of integer values
+ * + various functions for collecting data statistic
+ * and for specific data analysis of RDF triples
*
*/
@@ -43,7 +39,7 @@ typedef struct _hashmap_element{
int* key;
int num;
int freq;
- int in_use;
+ char in_use;
any_t data;
} hashmap_element;
@@ -112,10 +108,26 @@ extern int hashmap_statistic_groupcs_by_
extern int hashmap_put(map_t in, int* key, int num, any_t value);
/*
+ * Add a pointer to the hashmap with some key
+ * This function is ONLY used for the case of
+ * the predicate in RDF triple
+ */
+extern int hashmap_put_forP(map_t in, int* key, int num, any_t value, int
support);
+
+/*
* Get an element from the hashmap. Return MAP_OK or MAP_MISSING.
*/
extern int hashmap_get(map_t in, int* key, int num, any_t *arg, char
isUpdateFreq, int *retfreq);
+
+/*
+ * Get your pointer out of the hashmap with a key
+ * This function is ONLY used for the case of
+ * the predicate in RDF triple
+ */
+
+extern int hashmap_get_forP(map_t in, int* key, any_t *arg);
+
/*
* Remove an element from the hashmap. Return MAP_OK or MAP_MISSING.
*/
diff --git a/monetdb5/extras/rdf/rdfschema.c b/monetdb5/extras/rdf/rdfschema.c
--- a/monetdb5/extras/rdf/rdfschema.c
+++ b/monetdb5/extras/rdf/rdfschema.c
@@ -73,22 +73,21 @@ static void putCStoHash(map_t csmap, int
}
-static void putPtoHash(map_t pmap, int value, oid *poid){
+static void putPtoHash(map_t pmap, int value, oid *poid, int support){
oid *getPoid;
oid *putPoid;
int err;
int* pkey;
- int freq;
pkey = (int*) malloc(sizeof(int));
*pkey = value;
- if (hashmap_get(pmap, pkey, 1,(void**)(&getPoid),1, &freq) != MAP_OK){
+ if (hashmap_get_forP(pmap, pkey,(void**)(&getPoid)) != MAP_OK){
putPoid = malloc(sizeof(oid));
*putPoid = *poid;
- err = hashmap_put(pmap, pkey, 1, putPoid);
+ err = hashmap_put_forP(pmap, pkey, 1, putPoid, support);
assert(err == MAP_OK);
(*poid)++;
@@ -125,7 +124,7 @@ static void getStatisticCSsBySize(map_t
printf(" --- Number of CS per size (Max = %d)--- \n", maximumNumP);
for (i = 1; i <= maximumNumP; i++){
- printf("%d : %d \n", i, statCS[i]);
+ printf("%d : %d \n", i, statCS[i]);
}
free(statCS);
@@ -149,10 +148,10 @@ static void getStatisticCSsBySupports(ma
/* Output the result */
- if (isWriteToFile == 0){
+ if (isWriteToFile == 0){
printf(" --- Number of CS per support (Max = %d)--- \n",
maxSupport);
for (i = 1; i <= maxSupport; i++){
- printf("%d : %d \n", i, statCS[i]);
+ printf("%d : %d \n", i, statCS[i]);
}
}
else {
@@ -173,6 +172,8 @@ static void getStatisticCSsBySupports(ma
free(statCS);
}
+
+
/* Extract CS from SPO triples table */
str
RDFextractCS(int *ret, bat *sbatid, bat *pbatid){
@@ -214,8 +215,6 @@ RDFextractCS(int *ret, bat *sbatid, bat
if (numP > maxNumProp)
maxNumProp = numP;
-
-
}
curS = *bt;
curP = 0;
@@ -229,12 +228,12 @@ RDFextractCS(int *ret, bat *sbatid, bat
exit(-1);
}
- if (curP != *pbt){
+ if (curP != *pbt){ /* Multi values property */
buff[numP] = *pbt;
numP++;
curP = *pbt;
}
- //printf("Travel sbat at %d value: %d , for pbat: %d \n",
(int) p, (int) *bt, (int) *pbt);
+ //printf("Travel sbat at %d value: %d , for pbat: %d \n", (int)
p, (int) *bt, (int) *pbt);
}
/*put the last CS */
@@ -293,7 +292,8 @@ RDFextractPfromPSO(int *ret, bat *sbatid
bt = (oid *) BUNtloc(pi, p);
if (*bt != curP){
if (p != 0){ /* Not the first S */
- putPtoHash(pMap, *bt, &Poid);
+ putPtoHash(pMap, *bt, &Poid, supportP);
_______________________________________________
checkin-list mailing list
checkin-list@monetdb.org
http://mail.monetdb.org/mailman/listinfo/checkin-list
