Changeset: de96ad676406 for MonetDB
URL: https://dev.monetdb.org/hg/MonetDB/rev/de96ad676406
Modified Files:
gdk/gdk.h
gdk/gdk_align.c
gdk/gdk_bbp.c
gdk/gdk_string.c
Branch: ustr
Log Message:
Introduced tvkey property, set if the vheap does not contain duplicates.
diffs (125 lines):
diff --git a/gdk/gdk.h b/gdk/gdk.h
--- a/gdk/gdk.h
+++ b/gdk/gdk.h
@@ -398,6 +398,7 @@ typedef struct BAT {
/* see also comment near BATassertProps() for more information
* about the properties */
bool tkey:1; /* no duplicate values present */
+ bool tvkey:1; /* no duplicate values in tvheap */
bool tnonil:1; /* there are no nils in the column */
bool tnil:1; /* there is a nil in the column */
bool tsorted:1; /* column is sorted in ascending order */
@@ -603,6 +604,7 @@ typedef struct BATiter {
uint8_t shift;
int8_t type;
bool key:1,
+ vkey:1,
nonil:1,
nil:1,
sorted:1,
@@ -655,6 +657,7 @@ bat_iterator_nolock(BAT *b)
.maxpos = isview ? BUN_NONE : b->tmaxpos,
.unique_est = b->tunique_est,
.key = b->tkey,
+ .vkey = b->tvkey,
.nonil = b->tnonil,
.nil = b->tnil,
.sorted = b->tsorted,
diff --git a/gdk/gdk_align.c b/gdk/gdk_align.c
--- a/gdk/gdk_align.c
+++ b/gdk/gdk_align.c
@@ -170,6 +170,7 @@ VIEWcreate(oid seq, BAT *b, BUN l, BUN h
* accelerator data. We need copies because in case of a mark,
* we are going to override a column with a void. */
bn->tkey = bi.key;
+ bn->tvkey = bi.vkey;
bn->tseqbase = bi.tseq;
bn->tsorted = bi.sorted;
bn->trevsorted = bi.revsorted;
diff --git a/gdk/gdk_bbp.c b/gdk/gdk_bbp.c
--- a/gdk/gdk_bbp.c
+++ b/gdk/gdk_bbp.c
@@ -519,7 +519,7 @@ heapinit(BAT *b, const char *buf,
#endif
#endif
- if (properties & (bbpversion <= GDKLIBRARY_USTR ? ~0x1F81 : ~0x3F81)) {
+ if (properties & (bbpversion <= GDKLIBRARY_USTR ? ~0x1F81 : ~0x7F81)) {
TRC_CRITICAL(GDK, "unknown properties are set: incompatible
database on line %d of BBP.dir\n", lineno);
return -1;
}
@@ -560,6 +560,7 @@ heapinit(BAT *b, const char *buf,
b->tnil = (properties & 0x0800) != 0;
b->tascii = (properties & 0x1000) != 0;
b->ustr = (bbpversion <= GDKLIBRARY_USTR) ? false : (properties &
0x2000) != 0;
+ b->tvkey = (bbpversion <= GDKLIBRARY_USTR) ? false : (properties &
0x4000) != 0;
b->tnosorted = (BUN) nosorted;
b->tnorevsorted = (BUN) norevsorted;
b->tunique_est = 0.0;
@@ -2096,19 +2097,20 @@ heap_entry(FILE *fp, BATiter *bi, BUN si
free = 0;
}
- return fprintf(fp, " %s %d %d %d " BUNFMT " " BUNFMT " " BUNFMT " "
+ return fprintf(fp, " %s %" PRIu16 " %d %d " BUNFMT " " BUNFMT " "
BUNFMT " "
BUNFMT " " OIDFMT " %zu %" PRIu64" %" PRIu64,
bi->type >= 0 ? BATatoms[bi->type].name :
ATOMunknown_name(bi->type),
bi->width,
bi->type == TYPE_void || bi->vh != NULL || bi->ustr,
- (((unsigned short) bi->sorted << 0) |
- ((unsigned short) bi->revsorted << 7) |
- ((unsigned short) bi->key << 8) |
- ((unsigned short) BATtdensebi(bi) << 9) |
- ((unsigned short) bi->nonil << 10) |
- ((unsigned short) bi->nil << 11) |
- ((unsigned short) bi->ascii << 12) |
- ((unsigned short) bi->ustr << 13)),
+ (((uint16_t) bi->sorted << 0) |
+ ((uint16_t) bi->revsorted << 7) |
+ ((uint16_t) bi->key << 8) |
+ ((uint16_t) BATtdensebi(bi) << 9) |
+ ((uint16_t) bi->nonil << 10) |
+ ((uint16_t) bi->nil << 11) |
+ ((uint16_t) bi->ascii << 12) |
+ ((uint16_t) bi->ustr << 13) |
+ ((uint16_t) bi->vkey << 14)),
bi->nokey[0] >= size || bi->nokey[1] >= size ? 0 :
bi->nokey[0],
bi->nokey[0] >= size || bi->nokey[1] >= size ? 0 :
bi->nokey[1],
bi->nosorted >= size ? 0 : bi->nosorted,
diff --git a/gdk/gdk_string.c b/gdk/gdk_string.c
--- a/gdk/gdk_string.c
+++ b/gdk/gdk_string.c
@@ -249,6 +249,7 @@ ustrPut(BAT *b, var_t *dst, const char *
return (var_t) -1;
}
}
+ assert(ustrbat->tvkey);
BUN p = BUNfnd(ustrbat, v);
if (p == BUN_NONE) {
/* string does not yet occur in ustrbat */
@@ -278,6 +279,7 @@ ustrPut(BAT *b, var_t *dst, const char *
MT_rwlock_wrunlock(&ustrbat->thashlock);
ustrbat->tunique_est = (double) ustrbat->batCount;
ustrbat->tkey = true;
+ ustrbat->tvkey = true;
MT_lock_unset(&ustrbat->theaplock);
}
var_t d = VarHeapVal(ustrbat->theap->base, p, ustrbat->twidth);
@@ -403,6 +405,7 @@ strPut(BAT *b, var_t *dst, const void *V
#endif
h->dirty = true;
b->tascii = true;
+ b->tvkey = true;
}
off = strHash(v);
@@ -505,7 +508,8 @@ strPut(BAT *b, var_t *dst, const void *V
* string */
pos -= sizeof(stridx_t);
*(stridx_t *) (h->base + pos) = *bucket;
- }
+ } else if (*bucket != 0)
+ b->tvkey = false; /* we no longer know for sure */
*bucket = (stridx_t) pos; /* set bucket to the new string */
h->dirty = true;
_______________________________________________
checkin-list mailing list -- [email protected]
To unsubscribe send an email to [email protected]
