/*
* Validity test for an old-format hstore.
* 0 = not valid
* 1 = valid but with "slop" in the length
* 2 = exactly valid
*/
static int
hstoreValidOldFormat(HStore *hs)
{
int count = hs->size_;
HOldEntry *entries = (HOldEntry *) ARRPTR(hs);
int vsize;
int lastpos = 0;
int i;
if (hs->size_ & HS_FLAG_NEWVERSION)
return 0;
/* New format uses an HEntry for key and another for value */
StaticAssertStmt(sizeof(HOldEntry) == 2 * sizeof(HEntry),
"old hstore format is not upward-compatible");
if (count == 0)
return 2;
if (count > 0xFFFFFFF)
return 0;
if (CALCDATASIZE(count, 0) > VARSIZE(hs))
return 0;
if (entries[0].pos != 0)
return 0;
/* key length must be nondecreasing */
for (i = 1; i < count; ++i)
{
if (entries[i].keylen < entries[i - 1].keylen)
return 0;
}
/*
* entry position must be strictly increasing, except for the first entry
* (which can be ""=>"" and thus zero-length); and all entries must be
* properly contiguous
*/
for (i = 0; i < count; ++i)
{
if (entries[i].pos != lastpos)
return 0;
lastpos += (entries[i].keylen
+ ((entries[i].valisnull) ? 0 : entries[i].vallen));
}
vsize = CALCDATASIZE(count, lastpos);
if (vsize > VARSIZE(hs))
return 0;
if (vsize != VARSIZE(hs))
return 1;
return 2;
}
Datum
to_tsvector(PG_FUNCTION_ARGS)
{
text *in = PG_GETARG_TEXT_P(1);
PRSTEXT prs;
tsvector *out = NULL;
TSCfgInfo *cfg = findcfg(PG_GETARG_INT32(0));
prs.lenwords = 32;
prs.curwords = 0;
prs.pos = 0;
prs.words = (WORD *) palloc(sizeof(WORD) * prs.lenwords);
parsetext_v2(cfg, &prs, VARDATA(in), VARSIZE(in) - VARHDRSZ);
PG_FREE_IF_COPY(in, 1);
if (prs.curwords)
out = makevalue(&prs);
else
{
pfree(prs.words);
out = palloc(CALCDATASIZE(0, 0));
out->len = CALCDATASIZE(0, 0);
out->size = 0;
}
PG_RETURN_POINTER(out);
}
Datum
to_tsvector_byid(PG_FUNCTION_ARGS)
{
Oid cfgId = PG_GETARG_OID(0);
text *in = PG_GETARG_TEXT_P(1);
ParsedText prs;
TSVector out;
prs.lenwords = (VARSIZE(in) - VARHDRSZ) / 6; /* just estimation of
* word's number */
if (prs.lenwords == 0)
prs.lenwords = 2;
prs.curwords = 0;
prs.pos = 0;
prs.words = (ParsedWord *) palloc(sizeof(ParsedWord) * prs.lenwords);
parsetext(cfgId, &prs, VARDATA(in), VARSIZE(in) - VARHDRSZ);
PG_FREE_IF_COPY(in, 1);
if (prs.curwords)
out = make_tsvector(&prs);
else
{
pfree(prs.words);
out = palloc(CALCDATASIZE(0, 0));
SET_VARSIZE(out, CALCDATASIZE(0, 0));
out->size = 0;
}
PG_RETURN_POINTER(out);
}
Datum
hstore_in(PG_FUNCTION_ARGS)
{
HSParser state;
int4 len,
buflen,
i;
HStore *out;
HEntry *entries;
char *ptr;
state.begin = PG_GETARG_CSTRING(0);
parse_hstore(&state);
if (state.pcur == 0)
{
freeHSParse(&state);
len = CALCDATASIZE(0, 0);
out = palloc(len);
out->len = len;
out->size = 0;
PG_RETURN_POINTER(out);
}
state.pcur = uniquePairs(state.pairs, state.pcur, &buflen);
len = CALCDATASIZE(state.pcur, buflen);
out = palloc(len);
out->len = len;
out->size = state.pcur;
entries = ARRPTR(out);
ptr = STRPTR(out);
for (i = 0; i < out->size; i++)
{
entries[i].keylen = state.pairs[i].keylen;
entries[i].pos = ptr - STRPTR(out);
memcpy(ptr, state.pairs[i].key, state.pairs[i].keylen);
ptr += entries[i].keylen;
entries[i].valisnull = state.pairs[i].isnull;
if (entries[i].valisnull)
entries[i].vallen = 4; /* null */
else
{
entries[i].vallen = state.pairs[i].vallen;
memcpy(ptr, state.pairs[i].val, state.pairs[i].vallen);
ptr += entries[i].vallen;
}
}
freeHSParse(&state);
PG_RETURN_POINTER(out);
}
Datum
tconvert(PG_FUNCTION_ARGS)
{
text *key;
text *val = NULL;
int len;
HStore *out;
if (PG_ARGISNULL(0))
PG_RETURN_NULL();
key = PG_GETARG_TEXT_P(0);
if (PG_ARGISNULL(1))
len = CALCDATASIZE(1, VARSIZE(key));
else
{
val = PG_GETARG_TEXT_P(1);
len = CALCDATASIZE(1, VARSIZE(key) + VARSIZE(val) - 2 * VARHDRSZ);
}
out = palloc(len);
SET_VARSIZE(out, len);
out->size = 1;
ARRPTR(out)->keylen = hstoreCheckKeyLen(VARSIZE(key) - VARHDRSZ);
if (PG_ARGISNULL(1))
{
ARRPTR(out)->vallen = 0;
ARRPTR(out)->valisnull = true;
}
else
{
ARRPTR(out)->vallen = hstoreCheckValLen(VARSIZE(val) - VARHDRSZ);
ARRPTR(out)->valisnull = false;
}
ARRPTR(out)->pos = 0;
memcpy(STRPTR(out), VARDATA(key), ARRPTR(out)->keylen);
if (!PG_ARGISNULL(1))
{
memcpy(STRPTR(out) + ARRPTR(out)->keylen, VARDATA(val), ARRPTR(out)->vallen);
PG_FREE_IF_COPY(val, 1);
}
PG_FREE_IF_COPY(key, 0);
PG_RETURN_POINTER(out);
}
Datum
tsvector_strip(PG_FUNCTION_ARGS)
{
TSVector in = PG_GETARG_TSVECTOR(0);
TSVector out;
int i,
len = 0;
WordEntry *arrin = ARRPTR(in),
*arrout;
char *cur;
for (i = 0; i < in->size; i++)
len += arrin[i].len;
len = CALCDATASIZE(in->size, len);
out = (TSVector) palloc0(len);
SET_VARSIZE(out, len);
out->size = in->size;
arrout = ARRPTR(out);
cur = STRPTR(out);
for (i = 0; i < in->size; i++)
{
memcpy(cur, STRPTR(in) + arrin[i].pos, arrin[i].len);
arrout[i].haspos = 0;
arrout[i].len = arrin[i].len;
arrout[i].pos = cur - STRPTR(out);
cur += arrout[i].len;
}
PG_FREE_IF_COPY(in, 0);
PG_RETURN_POINTER(out);
}
HStore *
hstorePairs(Pairs *pairs, int4 pcount, int4 buflen)
{
HStore *out;
HEntry *entry;
char *ptr;
char *buf;
int4 len;
int4 i;
len = CALCDATASIZE(pcount, buflen);
out = palloc(len);
SET_VARSIZE(out, len);
HS_SETCOUNT(out, pcount);
if (pcount == 0)
return out;
entry = ARRPTR(out);
buf = ptr = STRPTR(out);
for (i = 0; i < pcount; i++)
HS_ADDITEM(entry, buf, ptr, pairs[i]);
HS_FINALIZE(out, pcount, buf, ptr);
return out;
}
Datum
hstore_out(PG_FUNCTION_ARGS)
{
HStore *in = PG_GETARG_HS(0);
int buflen,
i;
char *out,
*ptr;
char *base = STRPTR(in);
HEntry *entries = ARRPTR(in);
if (in->size == 0)
{
out = palloc(1);
*out = '\0';
PG_FREE_IF_COPY(in, 0);
PG_RETURN_CSTRING(out);
}
buflen = (4 /* " */ + 2 /* => */ + 2 /* , */ ) * in->size +
2 /* esc */ * (in->len - CALCDATASIZE(in->size, 0));
out = ptr = palloc(buflen);
for (i = 0; i < in->size; i++)
{
*ptr++ = '"';
ptr = cpw(ptr, base + entries[i].pos, entries[i].keylen);
*ptr++ = '"';
*ptr++ = '=';
*ptr++ = '>';
if (entries[i].valisnull)
{
*ptr++ = 'N';
*ptr++ = 'U';
*ptr++ = 'L';
*ptr++ = 'L';
}
else
{
*ptr++ = '"';
ptr = cpw(ptr, base + entries[i].pos + entries[i].keylen, entries[i].vallen);
*ptr++ = '"';
}
if (i + 1 != in->size)
{
*ptr++ = ',';
*ptr++ = ' ';
}
}
*ptr = '\0';
PG_FREE_IF_COPY(in, 0);
PG_RETURN_CSTRING(out);
}
/*
* Validity test for a new-format hstore.
* 0 = not valid
* 1 = valid but with "slop" in the length
* 2 = exactly valid
*/
static int
hstoreValidNewFormat(HStore *hs)
{
int count = HS_COUNT(hs);
HEntry *entries = ARRPTR(hs);
int buflen = (count) ? HSE_ENDPOS(entries[2 * (count) - 1]) : 0;
int vsize = CALCDATASIZE(count, buflen);
int i;
if (hs->size_ & HS_FLAG_NEWVERSION)
return 2;
if (count == 0)
return 2;
if (!HSE_ISFIRST(entries[0]))
return 0;
if (vsize > VARSIZE(hs))
return 0;
/* entry position must be nondecreasing */
for (i = 1; i < 2 * count; ++i)
{
if (HSE_ISFIRST(entries[i])
|| (HSE_ENDPOS(entries[i]) < HSE_ENDPOS(entries[i - 1])))
return 0;
}
/* key length must be nondecreasing and keys must not be null */
for (i = 1; i < count; ++i)
{
if (HS_KEYLEN(entries, i) < HS_KEYLEN(entries, i - 1))
return 0;
if (HSE_ISNULL(entries[2 * i]))
return 0;
}
if (vsize != VARSIZE(hs))
return 1;
return 2;
}
Datum
hstore_hash(PG_FUNCTION_ARGS)
{
HStore *hs = PG_GETARG_HS(0);
Datum hval = hash_any((unsigned char *) VARDATA(hs),
VARSIZE(hs) - VARHDRSZ);
/*
* this is the only place in the code that cares whether the overall
* varlena size exactly matches the true data size; this assertion should
* be maintained by all the other code, but we make it explicit here.
*/
Assert(VARSIZE(hs) ==
(HS_COUNT(hs) != 0 ?
CALCDATASIZE(HS_COUNT(hs),
HSE_ENDPOS(ARRPTR(hs)[2 * HS_COUNT(hs) - 1])) :
HSHRDSIZE));
PG_FREE_IF_COPY(hs, 0);
PG_RETURN_DATUM(hval);
}
/*
* Trigger
*/
Datum
tsearch2(PG_FUNCTION_ARGS)
{
TriggerData *trigdata;
Trigger *trigger;
Relation rel;
HeapTuple rettuple = NULL;
TSCfgInfo *cfg = findcfg(get_currcfg());
int numidxattr,
i;
PRSTEXT prs;
Datum datum = (Datum) 0;
Oid funcoid = InvalidOid;
if (!CALLED_AS_TRIGGER(fcinfo))
/* internal error */
elog(ERROR, "TSearch: Not fired by trigger manager");
trigdata = (TriggerData *) fcinfo->context;
if (TRIGGER_FIRED_FOR_STATEMENT(trigdata->tg_event))
/* internal error */
elog(ERROR, "TSearch: Can't process STATEMENT events");
if (TRIGGER_FIRED_AFTER(trigdata->tg_event))
/* internal error */
elog(ERROR, "TSearch: Must be fired BEFORE event");
if (TRIGGER_FIRED_BY_INSERT(trigdata->tg_event))
rettuple = trigdata->tg_trigtuple;
else if (TRIGGER_FIRED_BY_UPDATE(trigdata->tg_event))
rettuple = trigdata->tg_newtuple;
else
/* internal error */
elog(ERROR, "TSearch: Unknown event");
trigger = trigdata->tg_trigger;
rel = trigdata->tg_relation;
if (trigger->tgnargs < 2)
/* internal error */
elog(ERROR, "TSearch: format tsearch2(tsvector_field, text_field1,...)");
numidxattr = SPI_fnumber(rel->rd_att, trigger->tgargs[0]);
if (numidxattr == SPI_ERROR_NOATTRIBUTE)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_COLUMN),
errmsg("tsvector column \"%s\" does not exist",
trigger->tgargs[0])));
prs.lenwords = 32;
prs.curwords = 0;
prs.pos = 0;
prs.words = (WORD *) palloc(sizeof(WORD) * prs.lenwords);
/* find all words in indexable column */
for (i = 1; i < trigger->tgnargs; i++)
{
int numattr;
Oid oidtype;
Datum txt_toasted;
bool isnull;
text *txt;
numattr = SPI_fnumber(rel->rd_att, trigger->tgargs[i]);
if (numattr == SPI_ERROR_NOATTRIBUTE)
{
funcoid = findFunc(trigger->tgargs[i]);
if (funcoid == InvalidOid)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_COLUMN),
errmsg("could not find function or field \"%s\"",
trigger->tgargs[i])));
continue;
}
oidtype = SPI_gettypeid(rel->rd_att, numattr);
/* We assume char() and varchar() are binary-equivalent to text */
if (!(oidtype == TEXTOID ||
oidtype == VARCHAROID ||
oidtype == BPCHAROID))
{
elog(WARNING, "TSearch: '%s' is not of character type",
trigger->tgargs[i]);
continue;
}
txt_toasted = SPI_getbinval(rettuple, rel->rd_att, numattr, &isnull);
if (isnull)
continue;
if (funcoid != InvalidOid)
{
text *txttmp = (text *) DatumGetPointer(OidFunctionCall1(
funcoid,
PointerGetDatum(txt_toasted)
));
txt = (text *) DatumGetPointer(PG_DETOAST_DATUM(PointerGetDatum(txttmp)));
if (txt == txttmp)
txt_toasted = PointerGetDatum(txt);
}
else
txt = (text *) DatumGetPointer(PG_DETOAST_DATUM(PointerGetDatum(txt_toasted)));
parsetext_v2(cfg, &prs, VARDATA(txt), VARSIZE(txt) - VARHDRSZ);
if (txt != (text *) DatumGetPointer(txt_toasted))
pfree(txt);
}
/* make tsvector value */
if (prs.curwords)
{
datum = PointerGetDatum(makevalue(&prs));
rettuple = SPI_modifytuple(rel, rettuple, 1, &numidxattr,
&datum, NULL);
pfree(DatumGetPointer(datum));
}
else
{
tsvector *out = palloc(CALCDATASIZE(0, 0));
out->len = CALCDATASIZE(0, 0);
out->size = 0;
datum = PointerGetDatum(out);
pfree(prs.words);
rettuple = SPI_modifytuple(rel, rettuple, 1, &numidxattr,
&datum, NULL);
}
if (rettuple == NULL)
/* internal error */
elog(ERROR, "TSearch: %d returned by SPI_modifytuple", SPI_result);
return PointerGetDatum(rettuple);
}
static Datum
tsvector_update_trigger(PG_FUNCTION_ARGS, bool config_column)
{
TriggerData *trigdata;
Trigger *trigger;
Relation rel;
HeapTuple rettuple = NULL;
int tsvector_attr_num,
i;
ParsedText prs;
Datum datum;
bool isnull;
text *txt;
Oid cfgId;
/* Check call context */
if (!CALLED_AS_TRIGGER(fcinfo)) /* internal error */
elog(ERROR, "tsvector_update_trigger: not fired by trigger manager");
trigdata = (TriggerData *) fcinfo->context;
if (!TRIGGER_FIRED_FOR_ROW(trigdata->tg_event))
elog(ERROR, "tsvector_update_trigger: must be fired for row");
if (!TRIGGER_FIRED_BEFORE(trigdata->tg_event))
elog(ERROR, "tsvector_update_trigger: must be fired BEFORE event");
if (TRIGGER_FIRED_BY_INSERT(trigdata->tg_event))
rettuple = trigdata->tg_trigtuple;
else if (TRIGGER_FIRED_BY_UPDATE(trigdata->tg_event))
rettuple = trigdata->tg_newtuple;
else
elog(ERROR, "tsvector_update_trigger: must be fired for INSERT or UPDATE");
trigger = trigdata->tg_trigger;
rel = trigdata->tg_relation;
if (trigger->tgnargs < 3)
elog(ERROR, "tsvector_update_trigger: arguments must be tsvector_field, ts_config, text_field1, ...)");
/* Find the target tsvector column */
tsvector_attr_num = SPI_fnumber(rel->rd_att, trigger->tgargs[0]);
if (tsvector_attr_num == SPI_ERROR_NOATTRIBUTE)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_COLUMN),
errmsg("tsvector column \"%s\" does not exist",
trigger->tgargs[0])));
if (!IsBinaryCoercible(SPI_gettypeid(rel->rd_att, tsvector_attr_num),
TSVECTOROID))
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
errmsg("column \"%s\" is not of tsvector type",
trigger->tgargs[0])));
/* Find the configuration to use */
if (config_column)
{
int config_attr_num;
config_attr_num = SPI_fnumber(rel->rd_att, trigger->tgargs[1]);
if (config_attr_num == SPI_ERROR_NOATTRIBUTE)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_COLUMN),
errmsg("configuration column \"%s\" does not exist",
trigger->tgargs[1])));
if (!IsBinaryCoercible(SPI_gettypeid(rel->rd_att, config_attr_num),
REGCONFIGOID))
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
errmsg("column \"%s\" is not of regconfig type",
trigger->tgargs[1])));
datum = SPI_getbinval(rettuple, rel->rd_att, config_attr_num, &isnull);
if (isnull)
ereport(ERROR,
(errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED),
errmsg("configuration column \"%s\" must not be null",
trigger->tgargs[1])));
cfgId = DatumGetObjectId(datum);
}
else
{
List *names;
names = stringToQualifiedNameList(trigger->tgargs[1]);
/* require a schema so that results are not search path dependent */
if (list_length(names) < 2)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("text search configuration name \"%s\" must be schema-qualified",
trigger->tgargs[1])));
cfgId = get_ts_config_oid(names, false);
}
/* initialize parse state */
prs.lenwords = 32;
prs.curwords = 0;
prs.pos = 0;
prs.words = (ParsedWord *) palloc(sizeof(ParsedWord) * prs.lenwords);
/* find all words in indexable column(s) */
for (i = 2; i < trigger->tgnargs; i++)
{
int numattr;
numattr = SPI_fnumber(rel->rd_att, trigger->tgargs[i]);
if (numattr == SPI_ERROR_NOATTRIBUTE)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_COLUMN),
errmsg("column \"%s\" does not exist",
trigger->tgargs[i])));
if (!IsBinaryCoercible(SPI_gettypeid(rel->rd_att, numattr), TEXTOID))
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
errmsg("column \"%s\" is not of a character type",
trigger->tgargs[i])));
datum = SPI_getbinval(rettuple, rel->rd_att, numattr, &isnull);
if (isnull)
continue;
txt = DatumGetTextP(datum);
parsetext(cfgId, &prs, VARDATA(txt), VARSIZE(txt) - VARHDRSZ);
if (txt != (text *) DatumGetPointer(datum))
pfree(txt);
}
/* make tsvector value */
if (prs.curwords)
{
datum = PointerGetDatum(make_tsvector(&prs));
rettuple = SPI_modifytuple(rel, rettuple, 1, &tsvector_attr_num,
&datum, NULL);
pfree(DatumGetPointer(datum));
}
else
{
TSVector out = palloc(CALCDATASIZE(0, 0));
SET_VARSIZE(out, CALCDATASIZE(0, 0));
out->size = 0;
datum = PointerGetDatum(out);
rettuple = SPI_modifytuple(rel, rettuple, 1, &tsvector_attr_num,
&datum, NULL);
pfree(prs.words);
}
if (rettuple == NULL) /* internal error */
elog(ERROR, "tsvector_update_trigger: %d returned by SPI_modifytuple",
SPI_result);
return PointerGetDatum(rettuple);
}
Datum
tsvector_concat(PG_FUNCTION_ARGS)
{
TSVector in1 = PG_GETARG_TSVECTOR(0);
TSVector in2 = PG_GETARG_TSVECTOR(1);
TSVector out;
WordEntry *ptr;
WordEntry *ptr1,
*ptr2;
WordEntryPos *p;
int maxpos = 0,
i,
j,
i1,
i2,
dataoff;
char *data,
*data1,
*data2;
ptr = ARRPTR(in1);
i = in1->size;
while (i--)
{
if ((j = POSDATALEN(in1, ptr)) != 0)
{
p = POSDATAPTR(in1, ptr);
while (j--)
{
if (WEP_GETPOS(*p) > maxpos)
maxpos = WEP_GETPOS(*p);
p++;
}
}
ptr++;
}
ptr1 = ARRPTR(in1);
ptr2 = ARRPTR(in2);
data1 = STRPTR(in1);
data2 = STRPTR(in2);
i1 = in1->size;
i2 = in2->size;
/* conservative estimate of space needed */
out = (TSVector) palloc0(VARSIZE(in1) + VARSIZE(in2));
SET_VARSIZE(out, VARSIZE(in1) + VARSIZE(in2));
out->size = in1->size + in2->size;
ptr = ARRPTR(out);
data = STRPTR(out);
dataoff = 0;
while (i1 && i2)
{
int cmp = compareEntry(data1, ptr1, data2, ptr2);
if (cmp < 0)
{ /* in1 first */
ptr->haspos = ptr1->haspos;
ptr->len = ptr1->len;
memcpy(data + dataoff, data1 + ptr1->pos, ptr1->len);
ptr->pos = dataoff;
dataoff += ptr1->len;
if (ptr->haspos)
{
dataoff = SHORTALIGN(dataoff);
memcpy(data + dataoff, _POSVECPTR(in1, ptr1), POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16));
dataoff += POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16);
}
ptr++;
ptr1++;
i1--;
}
else if (cmp > 0)
{ /* in2 first */
ptr->haspos = ptr2->haspos;
ptr->len = ptr2->len;
memcpy(data + dataoff, data2 + ptr2->pos, ptr2->len);
ptr->pos = dataoff;
dataoff += ptr2->len;
if (ptr->haspos)
{
int addlen = add_pos(in2, ptr2, out, ptr, maxpos);
if (addlen == 0)
ptr->haspos = 0;
else
{
dataoff = SHORTALIGN(dataoff);
dataoff += addlen * sizeof(WordEntryPos) + sizeof(uint16);
}
}
ptr++;
ptr2++;
i2--;
}
else
{
ptr->haspos = ptr1->haspos | ptr2->haspos;
ptr->len = ptr1->len;
memcpy(data + dataoff, data1 + ptr1->pos, ptr1->len);
ptr->pos = dataoff;
dataoff += ptr1->len;
if (ptr->haspos)
{
if (ptr1->haspos)
{
dataoff = SHORTALIGN(dataoff);
memcpy(data + dataoff, _POSVECPTR(in1, ptr1), POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16));
dataoff += POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16);
if (ptr2->haspos)
dataoff += add_pos(in2, ptr2, out, ptr, maxpos) * sizeof(WordEntryPos);
}
else /* must have ptr2->haspos */
{
int addlen = add_pos(in2, ptr2, out, ptr, maxpos);
if (addlen == 0)
ptr->haspos = 0;
else
{
dataoff = SHORTALIGN(dataoff);
dataoff += addlen * sizeof(WordEntryPos) + sizeof(uint16);
}
}
}
ptr++;
ptr1++;
ptr2++;
i1--;
i2--;
}
}
while (i1)
{
ptr->haspos = ptr1->haspos;
ptr->len = ptr1->len;
memcpy(data + dataoff, data1 + ptr1->pos, ptr1->len);
ptr->pos = dataoff;
dataoff += ptr1->len;
if (ptr->haspos)
{
dataoff = SHORTALIGN(dataoff);
memcpy(data + dataoff, _POSVECPTR(in1, ptr1), POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16));
dataoff += POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16);
}
ptr++;
ptr1++;
i1--;
}
while (i2)
{
ptr->haspos = ptr2->haspos;
ptr->len = ptr2->len;
memcpy(data + dataoff, data2 + ptr2->pos, ptr2->len);
ptr->pos = dataoff;
dataoff += ptr2->len;
if (ptr->haspos)
{
int addlen = add_pos(in2, ptr2, out, ptr, maxpos);
if (addlen == 0)
ptr->haspos = 0;
else
{
dataoff = SHORTALIGN(dataoff);
dataoff += addlen * sizeof(WordEntryPos) + sizeof(uint16);
}
}
ptr++;
ptr2++;
i2--;
}
/*
* Instead of checking each offset individually, we check for overflow of
* pos fields once at the end.
*/
if (dataoff > MAXSTRPOS)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("string is too long for tsvector (%d bytes, max %d bytes)", dataoff, MAXSTRPOS)));
out->size = ptr - ARRPTR(out);
SET_VARSIZE(out, CALCDATASIZE(out->size, dataoff));
if (data != STRPTR(out))
memmove(STRPTR(out), data, dataoff);
PG_FREE_IF_COPY(in1, 0);
PG_FREE_IF_COPY(in2, 1);
PG_RETURN_POINTER(out);
}
Datum
tsvectorin(PG_FUNCTION_ARGS)
{
char *buf = PG_GETARG_CSTRING(0);
TSVectorParseState state;
WordEntryIN *arr;
int totallen;
int arrlen; /* allocated size of arr */
WordEntry *inarr;
int len = 0;
TSVector in;
int i;
char *token;
int toklen;
WordEntryPos *pos;
int poslen;
char *strbuf;
int stroff;
/*
* Tokens are appended to tmpbuf, cur is a pointer to the end of used
* space in tmpbuf.
*/
char *tmpbuf;
char *cur;
int buflen = 256; /* allocated size of tmpbuf */
state = init_tsvector_parser(buf, false, false);
arrlen = 64;
arr = (WordEntryIN *) palloc(sizeof(WordEntryIN) * arrlen);
cur = tmpbuf = (char *) palloc(buflen);
while (gettoken_tsvector(state, &token, &toklen, &pos, &poslen, NULL))
{
if (toklen >= MAXSTRLEN)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("word is too long (%ld bytes, max %ld bytes)",
(long) toklen,
(long) (MAXSTRLEN - 1))));
if (cur - tmpbuf > MAXSTRPOS)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("string is too long for tsvector (%ld bytes, max %ld bytes)",
(long) (cur - tmpbuf), (long) MAXSTRPOS)));
/*
* Enlarge buffers if needed
*/
if (len >= arrlen)
{
arrlen *= 2;
arr = (WordEntryIN *)
repalloc((void *) arr, sizeof(WordEntryIN) * arrlen);
}
while ((cur - tmpbuf) + toklen >= buflen)
{
int dist = cur - tmpbuf;
buflen *= 2;
tmpbuf = (char *) repalloc((void *) tmpbuf, buflen);
cur = tmpbuf + dist;
}
arr[len].entry.len = toklen;
arr[len].entry.pos = cur - tmpbuf;
memcpy((void *) cur, (void *) token, toklen);
cur += toklen;
if (poslen != 0)
{
arr[len].entry.haspos = 1;
arr[len].pos = pos;
arr[len].poslen = poslen;
}
else
{
arr[len].entry.haspos = 0;
arr[len].pos = NULL;
arr[len].poslen = 0;
}
len++;
}
close_tsvector_parser(state);
if (len > 0)
len = uniqueentry(arr, len, tmpbuf, &buflen);
else
buflen = 0;
if (buflen > MAXSTRPOS)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("string is too long for tsvector (%d bytes, max %d bytes)", buflen, MAXSTRPOS)));
totallen = CALCDATASIZE(len, buflen);
in = (TSVector) palloc0(totallen);
SET_VARSIZE(in, totallen);
in->size = len;
inarr = ARRPTR(in);
strbuf = STRPTR(in);
stroff = 0;
for (i = 0; i < len; i++)
{
memcpy(strbuf + stroff, &tmpbuf[arr[i].entry.pos], arr[i].entry.len);
arr[i].entry.pos = stroff;
stroff += arr[i].entry.len;
if (arr[i].entry.haspos)
{
if (arr[i].poslen > 0xFFFF)
elog(ERROR, "positions array too long");
/* Copy number of positions */
stroff = SHORTALIGN(stroff);
*(uint16 *) (strbuf + stroff) = (uint16) arr[i].poslen;
stroff += sizeof(uint16);
/* Copy positions */
memcpy(strbuf + stroff, arr[i].pos, arr[i].poslen * sizeof(WordEntryPos));
stroff += arr[i].poslen * sizeof(WordEntryPos);
pfree(arr[i].pos);
}
inarr[i] = arr[i].entry;
}
Assert((strbuf + stroff - (char *) in) == totallen);
PG_RETURN_TSVECTOR(in);
}
/*
* make value of tsvector, given parsed text
*/
TSVector
make_tsvector(ParsedText *prs)
{
int i,
j,
lenstr = 0,
totallen;
TSVector in;
WordEntry *ptr;
char *str;
int stroff;
prs->curwords = uniqueWORD(prs->words, prs->curwords);
for (i = 0; i < prs->curwords; i++)
{
lenstr += prs->words[i].len;
if (prs->words[i].alen)
{
lenstr = SHORTALIGN(lenstr);
lenstr += sizeof(uint16) + prs->words[i].pos.apos[0] * sizeof(WordEntryPos);
}
}
if (lenstr > MAXSTRPOS)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("string is too long for tsvector (%d bytes, max %d bytes)", lenstr, MAXSTRPOS)));
totallen = CALCDATASIZE(prs->curwords, lenstr);
in = (TSVector) palloc0(totallen);
SET_VARSIZE(in, totallen);
in->size = prs->curwords;
ptr = ARRPTR(in);
str = STRPTR(in);
stroff = 0;
for (i = 0; i < prs->curwords; i++)
{
ptr->len = prs->words[i].len;
ptr->pos = stroff;
memcpy(str + stroff, prs->words[i].word, prs->words[i].len);
stroff += prs->words[i].len;
pfree(prs->words[i].word);
if (prs->words[i].alen)
{
int k = prs->words[i].pos.apos[0];
WordEntryPos *wptr;
if (k > 0xFFFF)
elog(ERROR, "positions array too long");
ptr->haspos = 1;
stroff = SHORTALIGN(stroff);
*(uint16 *) (str + stroff) = (uint16) k;
wptr = POSDATAPTR(in, ptr);
for (j = 0; j < k; j++)
{
WEP_SETWEIGHT(wptr[j], 0);
WEP_SETPOS(wptr[j], prs->words[i].pos.apos[j + 1]);
}
stroff += sizeof(uint16) + k * sizeof(WordEntryPos);
pfree(prs->words[i].pos.apos);
}
else
ptr->haspos = 0;
ptr++;
}
pfree(prs->words);
return in;
}
Datum
hs_concat(PG_FUNCTION_ARGS)
{
HStore *s1 = PG_GETARG_HS(0);
HStore *s2 = PG_GETARG_HS(1);
HStore *out = palloc(VARSIZE(s1) + VARSIZE(s2));
char *ps1,
*ps2,
*pd;
HEntry *es1,
*es2,
*ed;
SET_VARSIZE(out, VARSIZE(s1) + VARSIZE(s2));
out->size = s1->size + s2->size;
ps1 = STRPTR(s1);
ps2 = STRPTR(s2);
pd = STRPTR(out);
es1 = ARRPTR(s1);
es2 = ARRPTR(s2);
ed = ARRPTR(out);
while (es1 - ARRPTR(s1) < s1->size && es2 - ARRPTR(s2) < s2->size)
{
int difference;
if (es1->keylen == es2->keylen)
difference = strncmp(ps1, ps2, es1->keylen);
else
difference = (es1->keylen > es2->keylen) ? 1 : -1;
if (difference == 0)
{
memcpy(ed, es2, sizeof(HEntry));
memcpy(pd, ps2, es2->keylen + ((es2->valisnull) ? 0 : es2->vallen));
ed->pos = pd - STRPTR(out);
pd += es2->keylen + ((es2->valisnull) ? 0 : es2->vallen);
ed++;
ps1 += es1->keylen + ((es1->valisnull) ? 0 : es1->vallen);
es1++;
ps2 += es2->keylen + ((es2->valisnull) ? 0 : es2->vallen);
es2++;
}
else if (difference > 0)
{
memcpy(ed, es2, sizeof(HEntry));
memcpy(pd, ps2, es2->keylen + ((es2->valisnull) ? 0 : es2->vallen));
ed->pos = pd - STRPTR(out);
pd += es2->keylen + ((es2->valisnull) ? 0 : es2->vallen);
ed++;
ps2 += es2->keylen + ((es2->valisnull) ? 0 : es2->vallen);
es2++;
}
else
{
memcpy(ed, es1, sizeof(HEntry));
memcpy(pd, ps1, es1->keylen + ((es1->valisnull) ? 0 : es1->vallen));
ed->pos = pd - STRPTR(out);
pd += es1->keylen + ((es1->valisnull) ? 0 : es1->vallen);
ed++;
ps1 += es1->keylen + ((es1->valisnull) ? 0 : es1->vallen);
es1++;
}
}
while (es1 - ARRPTR(s1) < s1->size)
{
memcpy(ed, es1, sizeof(HEntry));
memcpy(pd, ps1, es1->keylen + ((es1->valisnull) ? 0 : es1->vallen));
ed->pos = pd - STRPTR(out);
pd += es1->keylen + ((es1->valisnull) ? 0 : es1->vallen);
ed++;
ps1 += es1->keylen + ((es1->valisnull) ? 0 : es1->vallen);
es1++;
}
while (es2 - ARRPTR(s2) < s2->size)
{
memcpy(ed, es2, sizeof(HEntry));
memcpy(pd, ps2, es2->keylen + ((es2->valisnull) ? 0 : es2->vallen));
ed->pos = pd - STRPTR(out);
pd += es2->keylen + ((es2->valisnull) ? 0 : es2->vallen);
ed++;
ps2 += es2->keylen + ((es2->valisnull) ? 0 : es2->vallen);
es2++;
}
if (ed - ARRPTR(out) != out->size)
{
int buflen = pd - STRPTR(out);
pd = STRPTR(out);
out->size = ed - ARRPTR(out);
memmove(STRPTR(out), pd, buflen);
SET_VARSIZE(out, CALCDATASIZE(out->size, buflen));
}
PG_FREE_IF_COPY(s1, 0);
PG_FREE_IF_COPY(s2, 1);
PG_RETURN_POINTER(out);
}
/*
* make value of tsvector
*/
static tsvector *
makevalue(PRSTEXT * prs)
{
int4 i,
j,
lenstr = 0,
totallen;
tsvector *in;
WordEntry *ptr;
char *str,
*cur;
prs->curwords = uniqueWORD(prs->words, prs->curwords);
for (i = 0; i < prs->curwords; i++)
{
lenstr += SHORTALIGN(prs->words[i].len);
if (prs->words[i].alen)
lenstr += sizeof(uint16) + prs->words[i].pos.apos[0] * sizeof(WordEntryPos);
}
totallen = CALCDATASIZE(prs->curwords, lenstr);
in = (tsvector *) palloc(totallen);
memset(in, 0, totallen);
in->len = totallen;
in->size = prs->curwords;
ptr = ARRPTR(in);
cur = str = STRPTR(in);
for (i = 0; i < prs->curwords; i++)
{
ptr->len = prs->words[i].len;
if (cur - str > MAXSTRPOS)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("value is too big")));
ptr->pos = cur - str;
memcpy((void *) cur, (void *) prs->words[i].word, prs->words[i].len);
pfree(prs->words[i].word);
cur += SHORTALIGN(prs->words[i].len);
if (prs->words[i].alen)
{
WordEntryPos *wptr;
ptr->haspos = 1;
*(uint16 *) cur = prs->words[i].pos.apos[0];
wptr = POSDATAPTR(in, ptr);
for (j = 0; j < *(uint16 *) cur; j++)
{
wptr[j].weight = 0;
wptr[j].pos = prs->words[i].pos.apos[j + 1];
}
cur += sizeof(uint16) + prs->words[i].pos.apos[0] * sizeof(WordEntryPos);
pfree(prs->words[i].pos.apos);
}
else
ptr->haspos = 0;
ptr++;
}
pfree(prs->words);
return in;
}
Datum
tsvector_concat(PG_FUNCTION_ARGS)
{
TSVector in1 = PG_GETARG_TSVECTOR(0);
TSVector in2 = PG_GETARG_TSVECTOR(1);
TSVector out;
WordEntry *ptr;
WordEntry *ptr1,
*ptr2;
WordEntryPos *p;
int maxpos = 0,
i,
j,
i1,
i2,
dataoff,
output_bytes,
output_size;
char *data,
*data1,
*data2;
/* Get max position in in1; we'll need this to offset in2's positions */
ptr = ARRPTR(in1);
i = in1->size;
while (i--)
{
if ((j = POSDATALEN(in1, ptr)) != 0)
{
p = POSDATAPTR(in1, ptr);
while (j--)
{
if (WEP_GETPOS(*p) > maxpos)
maxpos = WEP_GETPOS(*p);
p++;
}
}
ptr++;
}
ptr1 = ARRPTR(in1);
ptr2 = ARRPTR(in2);
data1 = STRPTR(in1);
data2 = STRPTR(in2);
i1 = in1->size;
i2 = in2->size;
/*
* Conservative estimate of space needed. We might need all the data in
* both inputs, and conceivably add a pad byte before position data for
* each item where there was none before.
*/
output_bytes = VARSIZE(in1) + VARSIZE(in2) + i1 + i2;
out = (TSVector) palloc0(output_bytes);
SET_VARSIZE(out, output_bytes);
/*
* We must make out->size valid so that STRPTR(out) is sensible. We'll
* collapse out any unused space at the end.
*/
out->size = in1->size + in2->size;
ptr = ARRPTR(out);
data = STRPTR(out);
dataoff = 0;
while (i1 && i2)
{
int cmp = compareEntry(data1, ptr1, data2, ptr2);
if (cmp < 0)
{ /* in1 first */
ptr->haspos = ptr1->haspos;
ptr->len = ptr1->len;
memcpy(data + dataoff, data1 + ptr1->pos, ptr1->len);
ptr->pos = dataoff;
dataoff += ptr1->len;
if (ptr->haspos)
{
dataoff = SHORTALIGN(dataoff);
memcpy(data + dataoff, _POSVECPTR(in1, ptr1), POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16));
dataoff += POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16);
}
ptr++;
ptr1++;
i1--;
}
else if (cmp > 0)
{ /* in2 first */
ptr->haspos = ptr2->haspos;
ptr->len = ptr2->len;
memcpy(data + dataoff, data2 + ptr2->pos, ptr2->len);
ptr->pos = dataoff;
dataoff += ptr2->len;
if (ptr->haspos)
{
int addlen = add_pos(in2, ptr2, out, ptr, maxpos);
if (addlen == 0)
ptr->haspos = 0;
else
{
dataoff = SHORTALIGN(dataoff);
dataoff += addlen * sizeof(WordEntryPos) + sizeof(uint16);
}
}
ptr++;
ptr2++;
i2--;
}
else
{
ptr->haspos = ptr1->haspos | ptr2->haspos;
ptr->len = ptr1->len;
memcpy(data + dataoff, data1 + ptr1->pos, ptr1->len);
ptr->pos = dataoff;
dataoff += ptr1->len;
if (ptr->haspos)
{
if (ptr1->haspos)
{
dataoff = SHORTALIGN(dataoff);
memcpy(data + dataoff, _POSVECPTR(in1, ptr1), POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16));
dataoff += POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16);
if (ptr2->haspos)
dataoff += add_pos(in2, ptr2, out, ptr, maxpos) * sizeof(WordEntryPos);
}
else /* must have ptr2->haspos */
{
int addlen = add_pos(in2, ptr2, out, ptr, maxpos);
if (addlen == 0)
ptr->haspos = 0;
else
{
dataoff = SHORTALIGN(dataoff);
dataoff += addlen * sizeof(WordEntryPos) + sizeof(uint16);
}
}
}
ptr++;
ptr1++;
ptr2++;
i1--;
i2--;
}
}
while (i1)
{
ptr->haspos = ptr1->haspos;
ptr->len = ptr1->len;
memcpy(data + dataoff, data1 + ptr1->pos, ptr1->len);
ptr->pos = dataoff;
dataoff += ptr1->len;
if (ptr->haspos)
{
dataoff = SHORTALIGN(dataoff);
memcpy(data + dataoff, _POSVECPTR(in1, ptr1), POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16));
dataoff += POSDATALEN(in1, ptr1) * sizeof(WordEntryPos) + sizeof(uint16);
}
ptr++;
ptr1++;
i1--;
}
while (i2)
{
ptr->haspos = ptr2->haspos;
ptr->len = ptr2->len;
memcpy(data + dataoff, data2 + ptr2->pos, ptr2->len);
ptr->pos = dataoff;
dataoff += ptr2->len;
if (ptr->haspos)
{
int addlen = add_pos(in2, ptr2, out, ptr, maxpos);
if (addlen == 0)
ptr->haspos = 0;
else
{
dataoff = SHORTALIGN(dataoff);
dataoff += addlen * sizeof(WordEntryPos) + sizeof(uint16);
}
}
ptr++;
ptr2++;
i2--;
}
/*
* Instead of checking each offset individually, we check for overflow of
* pos fields once at the end.
*/
if (dataoff > MAXSTRPOS)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("string is too long for tsvector (%d bytes, max %d bytes)", dataoff, MAXSTRPOS)));
/*
* Adjust sizes (asserting that we didn't overrun the original estimates)
* and collapse out any unused array entries.
*/
output_size = ptr - ARRPTR(out);
Assert(output_size <= out->size);
out->size = output_size;
if (data != STRPTR(out))
memmove(STRPTR(out), data, dataoff);
output_bytes = CALCDATASIZE(out->size, dataoff);
Assert(output_bytes <= VARSIZE(out));
SET_VARSIZE(out, output_bytes);
PG_FREE_IF_COPY(in1, 0);
PG_FREE_IF_COPY(in2, 1);
PG_RETURN_POINTER(out);
}
Datum
tsvector_in(PG_FUNCTION_ARGS)
{
char *buf = PG_GETARG_CSTRING(0);
TI_IN_STATE state;
WordEntryIN *arr;
WordEntry *inarr;
int4 len = 0,
totallen = 64;
tsvector *in;
char *tmpbuf,
*cur;
int4 i,
buflen = 256;
state.prsbuf = buf;
state.len = 32;
state.word = (char *) palloc(state.len);
state.oprisdelim = false;
arr = (WordEntryIN *) palloc(sizeof(WordEntryIN) * totallen);
cur = tmpbuf = (char *) palloc(buflen);
while (gettoken_tsvector(&state))
{
if (len >= totallen)
{
totallen *= 2;
arr = (WordEntryIN *) repalloc((void *) arr, sizeof(WordEntryIN) * totallen);
}
while ((cur - tmpbuf) + (state.curpos - state.word) >= buflen)
{
int4 dist = cur - tmpbuf;
buflen *= 2;
tmpbuf = (char *) repalloc((void *) tmpbuf, buflen);
cur = tmpbuf + dist;
}
if (state.curpos - state.word >= MAXSTRLEN)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("word is too long")));
arr[len].entry.len = state.curpos - state.word;
if (cur - tmpbuf > MAXSTRPOS)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("too long value")));
arr[len].entry.pos = cur - tmpbuf;
memcpy((void *) cur, (void *) state.word, arr[len].entry.len);
cur += arr[len].entry.len;
if (state.alen)
{
arr[len].entry.haspos = 1;
arr[len].pos = state.pos;
}
else
arr[len].entry.haspos = 0;
len++;
}
pfree(state.word);
if (len > 0)
len = uniqueentry(arr, len, tmpbuf, &buflen);
else
buflen=0;
totallen = CALCDATASIZE(len, buflen);
in = (tsvector *) palloc(totallen);
memset(in, 0, totallen);
in->len = totallen;
in->size = len;
cur = STRPTR(in);
inarr = ARRPTR(in);
for (i = 0; i < len; i++)
{
memcpy((void *) cur, (void *) &tmpbuf[arr[i].entry.pos], arr[i].entry.len);
arr[i].entry.pos = cur - STRPTR(in);
cur += SHORTALIGN(arr[i].entry.len);
if (arr[i].entry.haspos)
{
memcpy(cur, arr[i].pos, (*(uint16 *) arr[i].pos + 1) * sizeof(WordEntryPos));
cur += (*(uint16 *) arr[i].pos + 1) * sizeof(WordEntryPos);
pfree(arr[i].pos);
}
memcpy(&(inarr[i]), &(arr[i].entry), sizeof(WordEntry));
}
pfree(tmpbuf);
pfree(arr);
PG_RETURN_POINTER(in);
}
