tsvector_op.c File Reference

#include "postgres.h"
#include <limits.h>
#include "access/htup_details.h"
#include "catalog/namespace.h"
#include "catalog/pg_type.h"
#include "commands/trigger.h"
#include "executor/spi.h"
#include "funcapi.h"
#include "mb/pg_wchar.h"
#include "miscadmin.h"
#include "parser/parse_coerce.h"
#include "tsearch/ts_utils.h"
#include "utils/builtins.h"
#include "utils/lsyscache.h"
#include "utils/regproc.h"
#include "utils/rel.h"

Include dependency graph for tsvector_op.c:

Go to the source code of this file.

Data Structures
struct	CHKVAL
struct	StatEntry
struct	TSVectorStat

Macros
#define	STATENTRYHDRSZ   (offsetof(StatEntry, lexeme))
#define	STATHDRSIZE   (offsetof(TSVectorStat, data))
#define	TSVECTORCMPFUNC(type, action, ret)
#define	compareEntry(pa, a, pb, b)
#define	TSPO_L_ONLY   0x01 /* emit positions appearing only in L */
#define	TSPO_R_ONLY   0x02 /* emit positions appearing only in R */
#define	TSPO_BOTH   0x04 /* emit positions appearing in both L&R */
#define	compareStatWord(a, e, t)

Typedefs
typedef struct StatEntry	StatEntry

Functions
static Datum	tsvector_update_trigger (PG_FUNCTION_ARGS, bool config_column)
static int	tsvector_bsearch (const TSVector tsv, char *lexeme, int lexeme_len)
static int	silly_cmp_tsvector (const TSVector a, const TSVector b)
	TSVECTORCMPFUNC (lt,<, BOOL)
	TSVECTORCMPFUNC (le,<=, BOOL)
	TSVECTORCMPFUNC (eq,==, BOOL)
	TSVECTORCMPFUNC (ge, >=, BOOL)
	TSVECTORCMPFUNC (gt, >, BOOL)
	TSVECTORCMPFUNC (ne,!=, BOOL)
	TSVECTORCMPFUNC (cmp,+, INT32)
Datum	tsvector_strip (PG_FUNCTION_ARGS)
Datum	tsvector_length (PG_FUNCTION_ARGS)
Datum	tsvector_setweight (PG_FUNCTION_ARGS)
Datum	tsvector_setweight_by_filter (PG_FUNCTION_ARGS)
static int32	add_pos (TSVector src, WordEntry *srcptr, TSVector dest, WordEntry *destptr, int32 maxpos)
static int	compare_int (const void *va, const void *vb)
static int	compare_text_lexemes (const void *va, const void *vb)
static TSVector	tsvector_delete_by_indices (TSVector tsv, int *indices_to_delete, int indices_count)
Datum	tsvector_delete_str (PG_FUNCTION_ARGS)
Datum	tsvector_delete_arr (PG_FUNCTION_ARGS)
Datum	tsvector_unnest (PG_FUNCTION_ARGS)
Datum	tsvector_to_array (PG_FUNCTION_ARGS)
Datum	array_to_tsvector (PG_FUNCTION_ARGS)
Datum	tsvector_filter (PG_FUNCTION_ARGS)
Datum	tsvector_concat (PG_FUNCTION_ARGS)
int32	tsCompareString (char *a, int lena, char *b, int lenb, bool prefix)
static bool	checkclass_str (CHKVAL *chkval, WordEntry *entry, QueryOperand *val, ExecPhraseData *data)
static int	uniqueLongPos (WordEntryPos *pos, int npos)
static bool	checkcondition_str (void *checkval, QueryOperand *val, ExecPhraseData *data)
static bool	TS_phrase_output (ExecPhraseData *data, ExecPhraseData *Ldata, ExecPhraseData *Rdata, int emit, int Loffset, int Roffset, int max_npos)
static bool	TS_phrase_execute (QueryItem *curitem, void *arg, uint32 flags, TSExecuteCallback chkcond, ExecPhraseData *data)
bool	TS_execute (QueryItem *curitem, void *arg, uint32 flags, TSExecuteCallback chkcond)
bool	tsquery_requires_match (QueryItem *curitem)
Datum	ts_match_qv (PG_FUNCTION_ARGS)
Datum	ts_match_vq (PG_FUNCTION_ARGS)
Datum	ts_match_tt (PG_FUNCTION_ARGS)
Datum	ts_match_tq (PG_FUNCTION_ARGS)
static int	check_weight (TSVector txt, WordEntry *wptr, int8 weight)
static void	insertStatEntry (MemoryContext persistentContext, TSVectorStat *stat, TSVector txt, uint32 off)
static void	chooseNextStatEntry (MemoryContext persistentContext, TSVectorStat *stat, TSVector txt, uint32 low, uint32 high, uint32 offset)
static TSVectorStat *	ts_accum (MemoryContext persistentContext, TSVectorStat *stat, Datum data)
static void	ts_setup_firstcall (FunctionCallInfo fcinfo, FuncCallContext *funcctx, TSVectorStat *stat)
static StatEntry *	walkStatEntryTree (TSVectorStat *stat)
static Datum	ts_process_call (FuncCallContext *funcctx)
static TSVectorStat *	ts_stat_sql (MemoryContext persistentContext, text *txt, text *ws)
Datum	ts_stat1 (PG_FUNCTION_ARGS)
Datum	ts_stat2 (PG_FUNCTION_ARGS)
Datum	tsvector_update_trigger_byid (PG_FUNCTION_ARGS)
Datum	tsvector_update_trigger_bycolumn (PG_FUNCTION_ARGS)

Macro Definition Documentation

#define compareEntry	(		pa,
			a,
			pb,
			b
	)

Value:

tsCompareString((pa) + (a)->pos, (a)->len,  \
                    (pb) + (b)->pos, (b)->len,  \
                    false)

Definition at line 347 of file tsvector_op.c.

Referenced by tsvector_concat().

#define compareStatWord	(		a,
			e,
			t
	)

Value:

tsCompareString((a)->lexeme, (a)->lenlexeme,        \
                    STRPTR(t) + (e)->pos, (e)->len,     \
                    false)

Definition at line 2019 of file tsvector_op.c.

Referenced by insertStatEntry().

#define STATENTRYHDRSZ   (offsetof(StatEntry, lexeme))

Definition at line 54 of file tsvector_op.c.

Referenced by insertStatEntry().

#define STATHDRSIZE   (offsetof(TSVectorStat, data))

Definition at line 68 of file tsvector_op.c.

#define TSPO_BOTH   0x04 /* emit positions appearing in both L&R */

Definition at line 1431 of file tsvector_op.c.

Referenced by TS_phrase_execute(), and TS_phrase_output().

#define TSPO_L_ONLY   0x01 /* emit positions appearing only in L */

Definition at line 1429 of file tsvector_op.c.

Referenced by TS_phrase_execute(), and TS_phrase_output().

#define TSPO_R_ONLY   0x02 /* emit positions appearing only in R */

Definition at line 1430 of file tsvector_op.c.

Referenced by TS_phrase_execute(), and TS_phrase_output().

#define TSVECTORCMPFUNC	(		type,
			action,
			ret
	)

Value:

Datum                                                   \
tsvector_##type(PG_FUNCTION_ARGS)                       \
{                                                       \
	TSVector   a = PG_GETARG_TSVECTOR(0);              \
	TSVector   b = PG_GETARG_TSVECTOR(1);              \
    int         res = silly_cmp_tsvector(a, b);         \
	PG_FREE_IF_COPY(a,0);                              \
	PG_FREE_IF_COPY(b,1);                              \
    PG_RETURN_##ret( res action 0 );                    \
}   \
/* keep compiler quiet - no extra ; */                  \
extern int no_such_variable

Definition at line 136 of file tsvector_op.c.

Typedef Documentation

typedef struct StatEntry StatEntry

Function Documentation

static int32 add_pos ( TSVector  src, WordEntry *  srcptr, TSVector  dest, WordEntry *  destptr, int32  maxpos  )

static

Definition at line 357 of file tsvector_op.c.

References _POSVECPTR, WordEntry::haspos, i, LIMITPOS, MAXENTRYPOS, MAXNUMPOS, POSDATALEN, POSDATAPTR, WEP_GETPOS, WEP_GETWEIGHT, WEP_SETPOS, and WEP_SETWEIGHT.

Referenced by tsvector_concat().

{
    uint16     *clen = &_POSVECPTR(dest, destptr)->npos;
    int         i;
    uint16      slen = POSDATALEN(src, srcptr),
                startlen;
    WordEntryPos *spos = POSDATAPTR(src, srcptr),
               *dpos = POSDATAPTR(dest, destptr);

    if (!destptr->haspos)
        *clen = 0;

    startlen = *clen;
    for (i = 0;
         i < slen && *clen < MAXNUMPOS &&
         (*clen == 0 || WEP_GETPOS(dpos[*clen - 1]) != MAXENTRYPOS - 1);
         i++)
    {
        WEP_SETWEIGHT(dpos[*clen], WEP_GETWEIGHT(spos[i]));
        WEP_SETPOS(dpos[*clen], LIMITPOS(WEP_GETPOS(spos[i]) + maxpos));
        (*clen)++;
    }

    if (*clen != startlen)
        destptr->haspos = 1;
    return *clen - startlen;
}

Datum array_to_tsvector

(

PG_FUNCTION_ARGS

)

Definition at line 756 of file tsvector_op.c.

References ARRPTR, CALCDATASIZE, compare_text_lexemes(), cur, deconstruct_array(), ereport, errcode(), errmsg(), ERROR, WordEntry::haspos, i, WordEntry::len, palloc0(), PG_FREE_IF_COPY, PG_GETARG_ARRAYTYPE_P, PG_RETURN_POINTER, WordEntry::pos, qsort, SET_VARSIZE, TSVectorData::size, STRPTR, TEXTOID, VARDATA, VARHDRSZ, and VARSIZE.

{
    ArrayType  *v = PG_GETARG_ARRAYTYPE_P(0);
    TSVector    tsout;
    Datum      *dlexemes;
    WordEntry  *arrout;
    bool       *nulls;
    int         nitems,
                i,
                j,
                tslen,
                datalen = 0;
    char       *cur;

    deconstruct_array(v, TEXTOID, -1, false, 'i', &dlexemes, &nulls, &nitems);

    /* Reject nulls (maybe we should just ignore them, instead?) */
    for (i = 0; i < nitems; i++)
    {
        if (nulls[i])
            ereport(ERROR,
                    (errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED),
                     errmsg("lexeme array may not contain nulls")));
    }

    /* Sort and de-dup, because this is required for a valid tsvector. */
    if (nitems > 1)
    {
        qsort(dlexemes, nitems, sizeof(Datum), compare_text_lexemes);
        j = 0;
        for (i = 1; i < nitems; i++)
        {
            if (compare_text_lexemes(&dlexemes[j], &dlexemes[i]) < 0)
                dlexemes[++j] = dlexemes[i];
        }
        nitems = ++j;
    }

    /* Calculate space needed for surviving lexemes. */
    for (i = 0; i < nitems; i++)
        datalen += VARSIZE(dlexemes[i]) - VARHDRSZ;
    tslen = CALCDATASIZE(nitems, datalen);

    /* Allocate and fill tsvector. */
    tsout = (TSVector) palloc0(tslen);
    SET_VARSIZE(tsout, tslen);
    tsout->size = nitems;

    arrout = ARRPTR(tsout);
    cur = STRPTR(tsout);
    for (i = 0; i < nitems; i++)
    {
        char       *lex = VARDATA(dlexemes[i]);
        int         lex_len = VARSIZE(dlexemes[i]) - VARHDRSZ;

        memcpy(cur, lex, lex_len);
        arrout[i].haspos = 0;
        arrout[i].len = lex_len;
        arrout[i].pos = cur - STRPTR(tsout);
        cur += lex_len;
    }

    PG_FREE_IF_COPY(v, 0);
    PG_RETURN_POINTER(tsout);
}

static int check_weight ( TSVector  txt, WordEntry *  wptr, int8  weight  )

static

Definition at line 2004 of file tsvector_op.c.

References POSDATALEN, POSDATAPTR, and WEP_GETWEIGHT.

Referenced by insertStatEntry().

{
    int         len = POSDATALEN(txt, wptr);
    int         num = 0;
    WordEntryPos *ptr = POSDATAPTR(txt, wptr);

    while (len--)
    {
        if (weight & (1 << WEP_GETWEIGHT(*ptr)))
            num++;
        ptr++;
    }
    return num;
}

static bool checkclass_str ( CHKVAL *  chkval, WordEntry *  entry, QueryOperand *  val, ExecPhraseData *  data  )

static

Definition at line 1197 of file tsvector_op.c.

References ExecPhraseData::allocated, WordEntry::haspos, WordEntry::len, WordEntryPosVector::npos, ExecPhraseData::npos, palloc(), WordEntry::pos, WordEntryPosVector::pos, ExecPhraseData::pos, result, SHORTALIGN, CHKVAL::values, QueryOperand::weight, WEP_GETPOS, and WEP_GETWEIGHT.

Referenced by checkcondition_str().

{
    bool        result = false;

    if (entry->haspos && (val->weight || data))
    {
        WordEntryPosVector *posvec;

        /*
         * We can't use the _POSVECPTR macro here because the pointer to the
         * tsvector's lexeme storage is already contained in chkval->values.
         */
        posvec = (WordEntryPosVector *)
            (chkval->values + SHORTALIGN(entry->pos + entry->len));

        if (val->weight && data)
        {
            WordEntryPos *posvec_iter = posvec->pos;
            WordEntryPos *dptr;

            /*
             * Filter position information by weights
             */
            dptr = data->pos = palloc(sizeof(WordEntryPos) * posvec->npos);
            data->allocated = true;

            /* Is there a position with a matching weight? */
            while (posvec_iter < posvec->pos + posvec->npos)
            {
                /* If true, append this position to the data->pos */
                if (val->weight & (1 << WEP_GETWEIGHT(*posvec_iter)))
                {
                    *dptr = WEP_GETPOS(*posvec_iter);
                    dptr++;
                }

                posvec_iter++;
            }

            data->npos = dptr - data->pos;

            if (data->npos > 0)
                result = true;
        }
        else if (val->weight)
        {
            WordEntryPos *posvec_iter = posvec->pos;

            /* Is there a position with a matching weight? */
            while (posvec_iter < posvec->pos + posvec->npos)
            {
                if (val->weight & (1 << WEP_GETWEIGHT(*posvec_iter)))
                {
                    result = true;
                    break;      /* no need to go further */
                }

                posvec_iter++;
            }
        }
        else                    /* data != NULL */
        {
            data->npos = posvec->npos;
            data->pos = posvec->pos;
            data->allocated = false;
            result = true;
        }
    }
    else
    {
        result = true;
    }

    return result;
}

static bool checkcondition_str ( void *  checkval, QueryOperand *  val, ExecPhraseData *  data  )

static

Definition at line 1311 of file tsvector_op.c.

References ExecPhraseData::allocated, CHKVAL::arrb, CHKVAL::arre, checkclass_str(), difference(), QueryOperand::distance, WordEntry::len, QueryOperand::length, ExecPhraseData::npos, CHKVAL::operand, palloc(), WordEntry::pos, ExecPhraseData::pos, QueryOperand::prefix, repalloc(), tsCompareString(), uniqueLongPos(), and CHKVAL::values.

Referenced by ts_match_vq().

{
    CHKVAL     *chkval = (CHKVAL *) checkval;
    WordEntry  *StopLow = chkval->arrb;
    WordEntry  *StopHigh = chkval->arre;
    WordEntry  *StopMiddle = StopHigh;
    int         difference = -1;
    bool        res = false;

    /* Loop invariant: StopLow <= val < StopHigh */
    while (StopLow < StopHigh)
    {
        StopMiddle = StopLow + (StopHigh - StopLow) / 2;
        difference = tsCompareString(chkval->operand + val->distance,
                                     val->length,
                                     chkval->values + StopMiddle->pos,
                                     StopMiddle->len,
                                     false);

        if (difference == 0)
        {
            /* Check weight info & fill 'data' with positions */
            res = checkclass_str(chkval, StopMiddle, val, data);
            break;
        }
        else if (difference > 0)
            StopLow = StopMiddle + 1;
        else
            StopHigh = StopMiddle;
    }

    if ((!res || data) && val->prefix)
    {
        WordEntryPos *allpos = NULL;
        int         npos = 0,
                    totalpos = 0;

        /*
         * there was a failed exact search, so we should scan further to find
         * a prefix match. We also need to do so if caller needs position info
         */
        if (StopLow >= StopHigh)
            StopMiddle = StopHigh;

        while ((!res || data) && StopMiddle < chkval->arre &&
               tsCompareString(chkval->operand + val->distance,
                               val->length,
                               chkval->values + StopMiddle->pos,
                               StopMiddle->len,
                               true) == 0)
        {
            if (data)
            {
                /*
                 * We need to join position information
                 */
                res = checkclass_str(chkval, StopMiddle, val, data);

                if (res)
                {
                    while (npos + data->npos >= totalpos)
                    {
                        if (totalpos == 0)
                        {
                            totalpos = 256;
                            allpos = palloc(sizeof(WordEntryPos) * totalpos);
                        }
                        else
                        {
                            totalpos *= 2;
                            allpos = repalloc(allpos, sizeof(WordEntryPos) * totalpos);
                        }
                    }

                    memcpy(allpos + npos, data->pos, sizeof(WordEntryPos) * data->npos);
                    npos += data->npos;
                }
            }
            else
            {
                res = checkclass_str(chkval, StopMiddle, val, NULL);
            }

            StopMiddle++;
        }

        if (res && data)
        {
            /* Sort and make unique array of found positions */
            data->pos = allpos;
            data->npos = uniqueLongPos(allpos, npos);
            data->allocated = true;
        }
    }

    return res;
}

static void chooseNextStatEntry ( MemoryContext  persistentContext, TSVectorStat *  stat, TSVector  txt, uint32  low, uint32  high, uint32  offset  )

static

Definition at line 2091 of file tsvector_op.c.

References insertStatEntry().

Referenced by ts_accum().

{
    uint32      pos;
    uint32      middle = (low + high) >> 1;

    pos = (low + middle) >> 1;
    if (low != middle && pos >= offset && pos - offset < txt->size)
        insertStatEntry(persistentContext, stat, txt, pos - offset);
    pos = (high + middle + 1) >> 1;
    if (middle + 1 != high && pos >= offset && pos - offset < txt->size)
        insertStatEntry(persistentContext, stat, txt, pos - offset);

    if (low != middle)
        chooseNextStatEntry(persistentContext, stat, txt, low, middle, offset);
    if (high != middle + 1)
        chooseNextStatEntry(persistentContext, stat, txt, middle + 1, high, offset);
}

static int compare_int ( const void *  va, const void *  vb  )

static

Definition at line 426 of file tsvector_op.c.

Referenced by tsvector_delete_by_indices().

{
    int         a = *((const int *) va);
    int         b = *((const int *) vb);

    if (a == b)
        return 0;
    return (a > b) ? 1 : -1;
}

static int compare_text_lexemes ( const void *  va, const void *  vb  )

static

Definition at line 437 of file tsvector_op.c.

References tsCompareString(), VARDATA_ANY, and VARSIZE_ANY_EXHDR.

Referenced by array_to_tsvector().

{
    Datum       a = *((const Datum *) va);
    Datum       b = *((const Datum *) vb);
    char       *alex = VARDATA_ANY(a);
    int         alex_len = VARSIZE_ANY_EXHDR(a);
    char       *blex = VARDATA_ANY(b);
    int         blex_len = VARSIZE_ANY_EXHDR(b);

    return tsCompareString(alex, alex_len, blex, blex_len, false);
}

static void insertStatEntry ( MemoryContext  persistentContext, TSVectorStat *  stat, TSVector  txt, uint32  off  )

static

Definition at line 2025 of file tsvector_op.c.

References ARRPTR, check_weight(), compareStatWord, WordEntry::haspos, StatEntry::left, WordEntry::len, StatEntry::lenlexeme, StatEntry::lexeme, TSVectorStat::maxdepth, MemoryContextAlloc(), StatEntry::ndoc, StatEntry::nentry, WordEntry::pos, POSDATALEN, StatEntry::right, TSVectorStat::root, STATENTRYHDRSZ, STRPTR, and TSVectorStat::weight.

Referenced by chooseNextStatEntry(), and ts_accum().

{
    WordEntry  *we = ARRPTR(txt) + off;
    StatEntry  *node = stat->root,
               *pnode = NULL;
    int         n,
                res = 0;
    uint32      depth = 1;

    if (stat->weight == 0)
        n = (we->haspos) ? POSDATALEN(txt, we) : 1;
    else
        n = (we->haspos) ? check_weight(txt, we, stat->weight) : 0;

    if (n == 0)
        return;                 /* nothing to insert */

    while (node)
    {
        res = compareStatWord(node, we, txt);

        if (res == 0)
        {
            break;
        }
        else
        {
            pnode = node;
            node = (res < 0) ? node->left : node->right;
        }
        depth++;
    }

    if (depth > stat->maxdepth)
        stat->maxdepth = depth;

    if (node == NULL)
    {
        node = MemoryContextAlloc(persistentContext, STATENTRYHDRSZ + we->len);
        node->left = node->right = NULL;
        node->ndoc = 1;
        node->nentry = n;
        node->lenlexeme = we->len;
        memcpy(node->lexeme, STRPTR(txt) + we->pos, node->lenlexeme);

        if (pnode == NULL)
        {
            stat->root = node;
        }
        else
        {
            if (res < 0)
                pnode->left = node;
            else
                pnode->right = node;
        }

    }
    else
    {
        node->ndoc++;
        node->nentry += n;
    }
}

static int silly_cmp_tsvector ( const TSVector  a, const TSVector  b  )

static

Definition at line 77 of file tsvector_op.c.

References ARRPTR, WordEntry::haspos, i, WordEntry::len, WordEntry::pos, POSDATALEN, POSDATAPTR, TSVectorData::size, STRPTR, tsCompareString(), VARSIZE, WEP_GETPOS, and WEP_GETWEIGHT.

{
    if (VARSIZE(a) < VARSIZE(b))
        return -1;
    else if (VARSIZE(a) > VARSIZE(b))
        return 1;
    else if (a->size < b->size)
        return -1;
    else if (a->size > b->size)
        return 1;
    else
    {
        WordEntry  *aptr = ARRPTR(a);
        WordEntry  *bptr = ARRPTR(b);
        int         i = 0;
        int         res;


        for (i = 0; i < a->size; i++)
        {
            if (aptr->haspos != bptr->haspos)
            {
                return (aptr->haspos > bptr->haspos) ? -1 : 1;
            }
            else if ((res = tsCompareString(STRPTR(a) + aptr->pos, aptr->len, STRPTR(b) + bptr->pos, bptr->len, false)) != 0)
            {
                return res;
            }
            else if (aptr->haspos)
            {
                WordEntryPos *ap = POSDATAPTR(a, aptr);
                WordEntryPos *bp = POSDATAPTR(b, bptr);
                int         j;

                if (POSDATALEN(a, aptr) != POSDATALEN(b, bptr))
                    return (POSDATALEN(a, aptr) > POSDATALEN(b, bptr)) ? -1 : 1;

                for (j = 0; j < POSDATALEN(a, aptr); j++)
                {
                    if (WEP_GETPOS(*ap) != WEP_GETPOS(*bp))
                    {
                        return (WEP_GETPOS(*ap) > WEP_GETPOS(*bp)) ? -1 : 1;
                    }
                    else if (WEP_GETWEIGHT(*ap) != WEP_GETWEIGHT(*bp))
                    {
                        return (WEP_GETWEIGHT(*ap) > WEP_GETWEIGHT(*bp)) ? -1 : 1;
                    }
                    ap++, bp++;
                }
            }

            aptr++;
            bptr++;
        }
    }

    return 0;
}

static TSVectorStat* ts_accum ( MemoryContext  persistentContext, TSVectorStat *  stat, Datum  data  )

static

Definition at line 2123 of file tsvector_op.c.

References chooseNextStatEntry(), DatumGetPointer, DatumGetTSVector, i, insertStatEntry(), TSVectorStat::maxdepth, MemoryContextAllocZero(), pfree(), and TSVectorData::size.

Referenced by ts_stat_sql().

{
    TSVector    txt = DatumGetTSVector(data);
    uint32      i,
                nbit = 0,
                offset;

    if (stat == NULL)
    {                           /* Init in first */
        stat = MemoryContextAllocZero(persistentContext, sizeof(TSVectorStat));
        stat->maxdepth = 1;
    }

    /* simple check of correctness */
    if (txt == NULL || txt->size == 0)
    {
        if (txt && txt != (TSVector) DatumGetPointer(data))
            pfree(txt);
        return stat;
    }

    i = txt->size - 1;
    for (; i > 0; i >>= 1)
        nbit++;

    nbit = 1 << nbit;
    offset = (nbit - txt->size) / 2;

    insertStatEntry(persistentContext, stat, txt, (nbit >> 1) - offset);
    chooseNextStatEntry(persistentContext, stat, txt, 0, nbit, offset);

    return stat;
}

bool TS_execute	(	QueryItem *	curitem,
		void *	arg,
		uint32	flags,
		TSExecuteCallback	chkcond
	)

Definition at line 1815 of file tsvector_op.c.

References check_stack_depth(), elog, ERROR, QueryOperator::left, OP_AND, OP_NOT, OP_OR, OP_PHRASE, QueryOperator::oper, QI_VAL, QueryItem::qoperator, TS_EXEC_CALC_NOT, TS_execute(), TS_phrase_execute(), and QueryItem::type.

Referenced by Cover(), gin_tsquery_consistent(), gtsvector_consistent(), hlCover(), TS_execute(), and ts_match_vq().

{
    /* since this function recurses, it could be driven to stack overflow */
    check_stack_depth();

    if (curitem->type == QI_VAL)
        return chkcond(arg, (QueryOperand *) curitem,
                       NULL /* we don't need position info */ );

    switch (curitem->qoperator.oper)
    {
        case OP_NOT:
            if (flags & TS_EXEC_CALC_NOT)
                return !TS_execute(curitem + 1, arg, flags, chkcond);
            else
                return true;

        case OP_AND:
            if (TS_execute(curitem + curitem->qoperator.left, arg, flags, chkcond))
                return TS_execute(curitem + 1, arg, flags, chkcond);
            else
                return false;

        case OP_OR:
            if (TS_execute(curitem + curitem->qoperator.left, arg, flags, chkcond))
                return true;
            else
                return TS_execute(curitem + 1, arg, flags, chkcond);

        case OP_PHRASE:
            return TS_phrase_execute(curitem, arg, flags, chkcond, NULL);

        default:
            elog(ERROR, "unrecognized operator: %d", curitem->qoperator.oper);
    }

    /* not reachable, but keep compiler quiet */
    return false;
}

Datum ts_match_qv

(

PG_FUNCTION_ARGS

)

Definition at line 1915 of file tsvector_op.c.

References DirectFunctionCall2, PG_GETARG_DATUM, PG_RETURN_DATUM, and ts_match_vq().

{
    PG_RETURN_DATUM(DirectFunctionCall2(ts_match_vq,
                                        PG_GETARG_DATUM(1),
                                        PG_GETARG_DATUM(0)));
}

Datum ts_match_tq

(

PG_FUNCTION_ARGS

)

Definition at line 1975 of file tsvector_op.c.

References DatumGetBool, DatumGetTSVector, DirectFunctionCall1, DirectFunctionCall2, pfree(), PG_FREE_IF_COPY, PG_GETARG_DATUM, PG_GETARG_TSQUERY, PG_RETURN_BOOL, to_tsvector(), ts_match_vq(), TSQueryGetDatum, and TSVectorGetDatum.

{
    TSVector    vector;
    TSQuery     query = PG_GETARG_TSQUERY(1);
    bool        res;

    vector = DatumGetTSVector(DirectFunctionCall1(to_tsvector,
                                                  PG_GETARG_DATUM(0)));

    res = DatumGetBool(DirectFunctionCall2(ts_match_vq,
                                           TSVectorGetDatum(vector),
                                           TSQueryGetDatum(query)));

    pfree(vector);
    PG_FREE_IF_COPY(query, 1);

    PG_RETURN_BOOL(res);
}

Datum ts_match_tt

(

PG_FUNCTION_ARGS

)

Definition at line 1953 of file tsvector_op.c.

References DatumGetBool, DatumGetTSQuery, DatumGetTSVector, DirectFunctionCall1, DirectFunctionCall2, pfree(), PG_GETARG_DATUM, PG_RETURN_BOOL, plainto_tsquery(), to_tsvector(), ts_match_vq(), TSQueryGetDatum, and TSVectorGetDatum.

{
    TSVector    vector;
    TSQuery     query;
    bool        res;

    vector = DatumGetTSVector(DirectFunctionCall1(to_tsvector,
                                                  PG_GETARG_DATUM(0)));
    query = DatumGetTSQuery(DirectFunctionCall1(plainto_tsquery,
                                                PG_GETARG_DATUM(1)));

    res = DatumGetBool(DirectFunctionCall2(ts_match_vq,
                                           TSVectorGetDatum(vector),
                                           TSQueryGetDatum(query)));

    pfree(vector);
    pfree(query);

    PG_RETURN_BOOL(res);
}

Datum ts_match_vq

(

PG_FUNCTION_ARGS

)

Definition at line 1923 of file tsvector_op.c.

References CHKVAL::arrb, CHKVAL::arre, ARRPTR, checkcondition_str(), GETOPERAND, GETQUERY, CHKVAL::operand, PG_FREE_IF_COPY, PG_GETARG_TSQUERY, PG_GETARG_TSVECTOR, PG_RETURN_BOOL, result, TSVectorData::size, TSQueryData::size, STRPTR, TS_EXEC_CALC_NOT, TS_execute(), val, and CHKVAL::values.

Referenced by ts_match_qv(), ts_match_tq(), and ts_match_tt().

{
    TSVector    val = PG_GETARG_TSVECTOR(0);
    TSQuery     query = PG_GETARG_TSQUERY(1);
    CHKVAL      chkval;
    bool        result;

    /* empty query matches nothing */
    if (!query->size)
    {
        PG_FREE_IF_COPY(val, 0);
        PG_FREE_IF_COPY(query, 1);
        PG_RETURN_BOOL(false);
    }

    chkval.arrb = ARRPTR(val);
    chkval.arre = chkval.arrb + val->size;
    chkval.values = STRPTR(val);
    chkval.operand = GETOPERAND(query);
    result = TS_execute(GETQUERY(query),
                        &chkval,
                        TS_EXEC_CALC_NOT,
                        checkcondition_str);

    PG_FREE_IF_COPY(val, 0);
    PG_FREE_IF_COPY(query, 1);
    PG_RETURN_BOOL(result);
}

static bool TS_phrase_execute ( QueryItem *  curitem, void *  arg, uint32  flags, TSExecuteCallback  chkcond, ExecPhraseData *  data  )

static

Definition at line 1572 of file tsvector_op.c.

References Assert, check_stack_depth(), QueryOperator::distance, elog, ERROR, QueryOperator::left, Max, Min, ExecPhraseData::negate, ExecPhraseData::npos, OP_AND, OP_NOT, OP_OR, OP_PHRASE, QueryOperator::oper, QI_VAL, QueryItem::qoperator, TS_EXEC_PHRASE_NO_POS, TS_phrase_output(), TSPO_BOTH, TSPO_L_ONLY, TSPO_R_ONLY, QueryItem::type, and ExecPhraseData::width.

Referenced by TS_execute().

{
    ExecPhraseData Ldata,
                Rdata;
    bool        lmatch,
                rmatch;
    int         Loffset,
                Roffset,
                maxwidth;

    /* since this function recurses, it could be driven to stack overflow */
    check_stack_depth();

    if (curitem->type == QI_VAL)
        return chkcond(arg, (QueryOperand *) curitem, data);

    switch (curitem->qoperator.oper)
    {
        case OP_NOT:

            /*
             * Because a "true" result with no specific positions is taken as
             * uncertain, we need no special care here for !TS_EXEC_CALC_NOT.
             * If it's a false positive, the right things happen anyway.
             *
             * Also, we need not touch data->width, since a NOT operation does
             * not change the match width.
             */
            if (TS_phrase_execute(curitem + 1, arg, flags, chkcond, data))
            {
                if (data->npos > 0)
                {
                    /* we have some positions, invert negate flag */
                    data->negate = !data->negate;
                    return true;
                }
                else if (data->negate)
                {
                    /* change "match everywhere" to "match nowhere" */
                    data->negate = false;
                    return false;
                }
                /* match positions are, and remain, uncertain */
                return true;
            }
            else
            {
                /* change "match nowhere" to "match everywhere" */
                Assert(data->npos == 0 && !data->negate);
                data->negate = true;
                return true;
            }

        case OP_PHRASE:
        case OP_AND:
            memset(&Ldata, 0, sizeof(Ldata));
            memset(&Rdata, 0, sizeof(Rdata));

            if (!TS_phrase_execute(curitem + curitem->qoperator.left,
                                   arg, flags, chkcond, &Ldata))
                return false;

            if (!TS_phrase_execute(curitem + 1,
                                   arg, flags, chkcond, &Rdata))
                return false;

            /*
             * If either operand has no position information, then we can't
             * return position data, only a "possible match" result. "Possible
             * match" answers are only wanted when TS_EXEC_PHRASE_NO_POS flag
             * is set, otherwise return false.
             */
            if ((Ldata.npos == 0 && !Ldata.negate) ||
                (Rdata.npos == 0 && !Rdata.negate))
                return (flags & TS_EXEC_PHRASE_NO_POS) ? true : false;

            if (curitem->qoperator.oper == OP_PHRASE)
            {
                /*
                 * Compute Loffset and Roffset suitable for phrase match, and
                 * compute overall width of whole phrase match.
                 */
                Loffset = curitem->qoperator.distance + Rdata.width;
                Roffset = 0;
                if (data)
                    data->width = curitem->qoperator.distance +
                        Ldata.width + Rdata.width;
            }
            else
            {
                /*
                 * For OP_AND, set output width and alignment like OP_OR (see
                 * comment below)
                 */
                maxwidth = Max(Ldata.width, Rdata.width);
                Loffset = maxwidth - Ldata.width;
                Roffset = maxwidth - Rdata.width;
                if (data)
                    data->width = maxwidth;
            }

            if (Ldata.negate && Rdata.negate)
            {
                /* !L & !R: treat as !(L | R) */
                (void) TS_phrase_output(data, &Ldata, &Rdata,
                                        TSPO_BOTH | TSPO_L_ONLY | TSPO_R_ONLY,
                                        Loffset, Roffset,
                                        Ldata.npos + Rdata.npos);
                if (data)
                    data->negate = true;
                return true;
            }
            else if (Ldata.negate)
            {
                /* !L & R */
                return TS_phrase_output(data, &Ldata, &Rdata,
                                        TSPO_R_ONLY,
                                        Loffset, Roffset,
                                        Rdata.npos);
            }
            else if (Rdata.negate)
            {
                /* L & !R */
                return TS_phrase_output(data, &Ldata, &Rdata,
                                        TSPO_L_ONLY,
                                        Loffset, Roffset,
                                        Ldata.npos);
            }
            else
            {
                /* straight AND */
                return TS_phrase_output(data, &Ldata, &Rdata,
                                        TSPO_BOTH,
                                        Loffset, Roffset,
                                        Min(Ldata.npos, Rdata.npos));
            }

        case OP_OR:
            memset(&Ldata, 0, sizeof(Ldata));
            memset(&Rdata, 0, sizeof(Rdata));

            lmatch = TS_phrase_execute(curitem + curitem->qoperator.left,
                                       arg, flags, chkcond, &Ldata);
            rmatch = TS_phrase_execute(curitem + 1,
                                       arg, flags, chkcond, &Rdata);

            if (!lmatch && !rmatch)
                return false;

            /*
             * If a valid operand has no position information, then we can't
             * return position data, only a "possible match" result. "Possible
             * match" answers are only wanted when TS_EXEC_PHRASE_NO_POS flag
             * is set, otherwise return false.
             */
            if ((lmatch && Ldata.npos == 0 && !Ldata.negate) ||
                (rmatch && Rdata.npos == 0 && !Rdata.negate))
                return (flags & TS_EXEC_PHRASE_NO_POS) ? true : false;

            /*
             * Cope with undefined output width from failed submatch.  (This
             * takes less code than trying to ensure that all failure returns
             * set data->width to zero.)
             */
            if (!lmatch)
                Ldata.width = 0;
            if (!rmatch)
                Rdata.width = 0;

            /*
             * For OP_AND and OP_OR, report the width of the wider of the two
             * inputs, and align the narrower input's positions to the right
             * end of that width.  This rule deals at least somewhat
             * reasonably with cases like "x <-> (y | z <-> q)".
             */
            maxwidth = Max(Ldata.width, Rdata.width);
            Loffset = maxwidth - Ldata.width;
            Roffset = maxwidth - Rdata.width;
            data->width = maxwidth;

            if (Ldata.negate && Rdata.negate)
            {
                /* !L | !R: treat as !(L & R) */
                (void) TS_phrase_output(data, &Ldata, &Rdata,
                                        TSPO_BOTH,
                                        Loffset, Roffset,
                                        Min(Ldata.npos, Rdata.npos));
                data->negate = true;
                return true;
            }
            else if (Ldata.negate)
            {
                /* !L | R: treat as !(L & !R) */
                (void) TS_phrase_output(data, &Ldata, &Rdata,
                                        TSPO_L_ONLY,
                                        Loffset, Roffset,
                                        Ldata.npos);
                data->negate = true;
                return true;
            }
            else if (Rdata.negate)
            {
                /* L | !R: treat as !(!L & R) */
                (void) TS_phrase_output(data, &Ldata, &Rdata,
                                        TSPO_R_ONLY,
                                        Loffset, Roffset,
                                        Rdata.npos);
                data->negate = true;
                return true;
            }
            else
            {
                /* straight OR */
                return TS_phrase_output(data, &Ldata, &Rdata,
                                        TSPO_BOTH | TSPO_L_ONLY | TSPO_R_ONLY,
                                        Loffset, Roffset,
                                        Ldata.npos + Rdata.npos);
            }

        default:
            elog(ERROR, "unrecognized operator: %d", curitem->qoperator.oper);
    }

    /* not reachable, but keep compiler quiet */
    return false;
}

static bool TS_phrase_output ( ExecPhraseData *  data, ExecPhraseData *  Ldata, ExecPhraseData *  Rdata, int  emit, int  Loffset, int  Roffset, int  max_npos  )

static

Definition at line 1434 of file tsvector_op.c.

References ExecPhraseData::allocated, Assert, ExecPhraseData::npos, palloc(), ExecPhraseData::pos, TSPO_BOTH, TSPO_L_ONLY, TSPO_R_ONLY, and WEP_GETPOS.

Referenced by TS_phrase_execute().

{
    int         Lindex,
                Rindex;

    /* Loop until both inputs are exhausted */
    Lindex = Rindex = 0;
    while (Lindex < Ldata->npos || Rindex < Rdata->npos)
    {
        int         Lpos,
                    Rpos;
        int         output_pos = 0;

        /*
         * Fetch current values to compare.  WEP_GETPOS() is needed because
         * ExecPhraseData->data can point to a tsvector's WordEntryPosVector.
         */
        if (Lindex < Ldata->npos)
            Lpos = WEP_GETPOS(Ldata->pos[Lindex]) + Loffset;
        else
        {
            /* L array exhausted, so we're done if R_ONLY isn't set */
            if (!(emit & TSPO_R_ONLY))
                break;
            Lpos = INT_MAX;
        }
        if (Rindex < Rdata->npos)
            Rpos = WEP_GETPOS(Rdata->pos[Rindex]) + Roffset;
        else
        {
            /* R array exhausted, so we're done if L_ONLY isn't set */
            if (!(emit & TSPO_L_ONLY))
                break;
            Rpos = INT_MAX;
        }

        /* Merge-join the two input lists */
        if (Lpos < Rpos)
        {
            /* Lpos is not matched in Rdata, should we output it? */
            if (emit & TSPO_L_ONLY)
                output_pos = Lpos;
            Lindex++;
        }
        else if (Lpos == Rpos)
        {
            /* Lpos and Rpos match ... should we output it? */
            if (emit & TSPO_BOTH)
                output_pos = Rpos;
            Lindex++;
            Rindex++;
        }
        else                    /* Lpos > Rpos */
        {
            /* Rpos is not matched in Ldata, should we output it? */
            if (emit & TSPO_R_ONLY)
                output_pos = Rpos;
            Rindex++;
        }

        if (output_pos > 0)
        {
            if (data)
            {
                /* Store position, first allocating output array if needed */
                if (data->pos == NULL)
                {
                    data->pos = (WordEntryPos *)
                        palloc(max_npos * sizeof(WordEntryPos));
                    data->allocated = true;
                }
                data->pos[data->npos++] = output_pos;
            }
            else
            {
                /*
                 * Exact positions not needed, so return true as soon as we
                 * know there is at least one.
                 */
                return true;
            }
        }
    }

    if (data && data->npos > 0)
    {
        /* Let's assert we didn't overrun the array */
        Assert(data->npos <= max_npos);
        return true;
    }
    return false;
}

static Datum ts_process_call ( FuncCallContext *  funcctx )

static

Definition at line 2250 of file tsvector_op.c.

References FuncCallContext::attinmeta, BuildTupleFromCStrings(), HeapTupleGetDatum, StatEntry::lenlexeme, StatEntry::lexeme, StatEntry::ndoc, StatEntry::nentry, palloc(), pfree(), result, FuncCallContext::user_fctx, values, and walkStatEntryTree().

Referenced by ts_stat1(), and ts_stat2().

{
    TSVectorStat *st;
    StatEntry  *entry;

    st = (TSVectorStat *) funcctx->user_fctx;

    entry = walkStatEntryTree(st);

    if (entry != NULL)
    {
        Datum       result;
        char       *values[3];
        char        ndoc[16];
        char        nentry[16];
        HeapTuple   tuple;

        values[0] = palloc(entry->lenlexeme + 1);
        memcpy(values[0], entry->lexeme, entry->lenlexeme);
        (values[0])[entry->lenlexeme] = '\0';
        sprintf(ndoc, "%d", entry->ndoc);
        values[1] = ndoc;
        sprintf(nentry, "%d", entry->nentry);
        values[2] = nentry;

        tuple = BuildTupleFromCStrings(funcctx->attinmeta, values);
        result = HeapTupleGetDatum(tuple);

        pfree(values[0]);

        /* mark entry as already visited */
        entry->ndoc = 0;

        return result;
    }

    return (Datum) 0;
}

static void ts_setup_firstcall ( FunctionCallInfo  fcinfo, FuncCallContext *  funcctx, TSVectorStat *  stat  )

static

Definition at line 2158 of file tsvector_op.c.

References Assert, FuncCallContext::attinmeta, BlessTupleDesc(), CreateTemplateTupleDesc(), INT4OID, StatEntry::left, TSVectorStat::maxdepth, MemoryContextSwitchTo(), FuncCallContext::multi_call_memory_ctx, palloc0(), TSVectorStat::root, TSVectorStat::stack, TSVectorStat::stackpos, TEXTOID, FuncCallContext::tuple_desc, TupleDescGetAttInMetadata(), TupleDescInitEntry(), and FuncCallContext::user_fctx.

Referenced by ts_stat1(), and ts_stat2().

{
    TupleDesc   tupdesc;
    MemoryContext oldcontext;
    StatEntry  *node;

    funcctx->user_fctx = (void *) stat;

    oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);

    stat->stack = palloc0(sizeof(StatEntry *) * (stat->maxdepth + 1));
    stat->stackpos = 0;

    node = stat->root;
    /* find leftmost value */
    if (node == NULL)
        stat->stack[stat->stackpos] = NULL;
    else
        for (;;)
        {
            stat->stack[stat->stackpos] = node;
            if (node->left)
            {
                stat->stackpos++;
                node = node->left;
            }
            else
                break;
        }
    Assert(stat->stackpos <= stat->maxdepth);

    tupdesc = CreateTemplateTupleDesc(3, false);
    TupleDescInitEntry(tupdesc, (AttrNumber) 1, "word",
                       TEXTOID, -1, 0);
    TupleDescInitEntry(tupdesc, (AttrNumber) 2, "ndoc",
                       INT4OID, -1, 0);
    TupleDescInitEntry(tupdesc, (AttrNumber) 3, "nentry",
                       INT4OID, -1, 0);
    funcctx->tuple_desc = BlessTupleDesc(tupdesc);
    funcctx->attinmeta = TupleDescGetAttInMetadata(tupdesc);

    MemoryContextSwitchTo(oldcontext);
}

Datum ts_stat1

(

PG_FUNCTION_ARGS

)

Definition at line 2379 of file tsvector_op.c.

References FuncCallContext::multi_call_memory_ctx, PG_FREE_IF_COPY, PG_GETARG_TEXT_PP, result, SPI_connect(), SPI_finish(), SRF_FIRSTCALL_INIT, SRF_IS_FIRSTCALL, SRF_PERCALL_SETUP, SRF_RETURN_DONE, SRF_RETURN_NEXT, ts_process_call(), ts_setup_firstcall(), and ts_stat_sql().

{
    FuncCallContext *funcctx;
    Datum       result;

    if (SRF_IS_FIRSTCALL())
    {
        TSVectorStat *stat;
        text       *txt = PG_GETARG_TEXT_PP(0);

        funcctx = SRF_FIRSTCALL_INIT();
        SPI_connect();
        stat = ts_stat_sql(funcctx->multi_call_memory_ctx, txt, NULL);
        PG_FREE_IF_COPY(txt, 0);
        ts_setup_firstcall(fcinfo, funcctx, stat);
        SPI_finish();
    }

    funcctx = SRF_PERCALL_SETUP();
    if ((result = ts_process_call(funcctx)) != (Datum) 0)
        SRF_RETURN_NEXT(funcctx, result);
    SRF_RETURN_DONE(funcctx);
}

Datum ts_stat2

(

PG_FUNCTION_ARGS

)

Definition at line 2404 of file tsvector_op.c.

References FuncCallContext::multi_call_memory_ctx, PG_FREE_IF_COPY, PG_GETARG_TEXT_PP, result, SPI_connect(), SPI_finish(), SRF_FIRSTCALL_INIT, SRF_IS_FIRSTCALL, SRF_PERCALL_SETUP, SRF_RETURN_DONE, SRF_RETURN_NEXT, ts_process_call(), ts_setup_firstcall(), and ts_stat_sql().

{
    FuncCallContext *funcctx;
    Datum       result;

    if (SRF_IS_FIRSTCALL())
    {
        TSVectorStat *stat;
        text       *txt = PG_GETARG_TEXT_PP(0);
        text       *ws = PG_GETARG_TEXT_PP(1);

        funcctx = SRF_FIRSTCALL_INIT();
        SPI_connect();
        stat = ts_stat_sql(funcctx->multi_call_memory_ctx, txt, ws);
        PG_FREE_IF_COPY(txt, 0);
        PG_FREE_IF_COPY(ws, 1);
        ts_setup_firstcall(fcinfo, funcctx, stat);
        SPI_finish();
    }

    funcctx = SRF_PERCALL_SETUP();
    if ((result = ts_process_call(funcctx)) != (Datum) 0)
        SRF_RETURN_NEXT(funcctx, result);
    SRF_RETURN_DONE(funcctx);
}

static TSVectorStat* ts_stat_sql ( MemoryContext  persistentContext, text *  txt, text *  ws  )

static

Definition at line 2290 of file tsvector_op.c.

References buf, elog, ereport, errcode(), errmsg(), ERROR, i, IsBinaryCoercible(), TSVectorStat::maxdepth, MemoryContextAllocZero(), tupleDesc::natts, pfree(), pg_mblen(), SPI_cursor_close(), SPI_cursor_fetch(), SPI_cursor_open(), SPI_freeplan(), SPI_freetuptable(), SPI_getbinval(), SPI_gettypeid(), SPI_prepare(), SPI_processed, SPI_tuptable, text_to_cstring(), ts_accum(), TSVECTOROID, SPITupleTable::tupdesc, SPITupleTable::vals, VARDATA_ANY, VARSIZE_ANY_EXHDR, and TSVectorStat::weight.

Referenced by ts_stat1(), and ts_stat2().

{
    char       *query = text_to_cstring(txt);
    TSVectorStat *stat;
    bool        isnull;
    Portal      portal;
    SPIPlanPtr  plan;

    if ((plan = SPI_prepare(query, 0, NULL)) == NULL)
        /* internal error */
        elog(ERROR, "SPI_prepare(\"%s\") failed", query);

    if ((portal = SPI_cursor_open(NULL, plan, NULL, NULL, true)) == NULL)
        /* internal error */
        elog(ERROR, "SPI_cursor_open(\"%s\") failed", query);

    SPI_cursor_fetch(portal, true, 100);

    if (SPI_tuptable == NULL ||
        SPI_tuptable->tupdesc->natts != 1 ||
        !IsBinaryCoercible(SPI_gettypeid(SPI_tuptable->tupdesc, 1),
                           TSVECTOROID))
        ereport(ERROR,
                (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                 errmsg("ts_stat query must return one tsvector column")));

    stat = MemoryContextAllocZero(persistentContext, sizeof(TSVectorStat));
    stat->maxdepth = 1;

    if (ws)
    {
        char       *buf;

        buf = VARDATA_ANY(ws);
        while (buf - VARDATA_ANY(ws) < VARSIZE_ANY_EXHDR(ws))
        {
            if (pg_mblen(buf) == 1)
            {
                switch (*buf)
                {
                    case 'A':
                    case 'a':
                        stat->weight |= 1 << 3;
                        break;
                    case 'B':
                    case 'b':
                        stat->weight |= 1 << 2;
                        break;
                    case 'C':
                    case 'c':
                        stat->weight |= 1 << 1;
                        break;
                    case 'D':
                    case 'd':
                        stat->weight |= 1;
                        break;
                    default:
                        stat->weight |= 0;
                }
            }
            buf += pg_mblen(buf);
        }
    }

    while (SPI_processed > 0)
    {
        uint64      i;

        for (i = 0; i < SPI_processed; i++)
        {
            Datum       data = SPI_getbinval(SPI_tuptable->vals[i], SPI_tuptable->tupdesc, 1, &isnull);

            if (!isnull)
                stat = ts_accum(persistentContext, stat, data);
        }

        SPI_freetuptable(SPI_tuptable);
        SPI_cursor_fetch(portal, true, 100);
    }

    SPI_freetuptable(SPI_tuptable);
    SPI_cursor_close(portal);
    SPI_freeplan(plan);
    pfree(query);

    return stat;
}

int32 tsCompareString	(	char *	a,
		int	lena,
		char *	b,
		int	lenb,
		bool	prefix
	)

Definition at line 1160 of file tsvector_op.c.

References cmp(), and Min.

Referenced by checkcondition_str(), compare_text_lexemes(), compareentry(), compareQueryOperand(), compareWORD(), gin_cmp_prefix(), gin_cmp_tslexeme(), hlfinditem(), QTNodeCompare(), silly_cmp_tsvector(), and tsvector_bsearch().

{
    int         cmp;

    if (lena == 0)
    {
        if (prefix)
            cmp = 0;            /* empty string is prefix of anything */
        else
            cmp = (lenb > 0) ? -1 : 0;
    }
    else if (lenb == 0)
    {
        cmp = (lena > 0) ? 1 : 0;
    }
    else
    {
        cmp = memcmp(a, b, Min(lena, lenb));

        if (prefix)
        {
            if (cmp == 0 && lena > lenb)
                cmp = 1;        /* a is longer, so not a prefix of b */
        }
        else if (cmp == 0 && lena != lenb)
        {
            cmp = (lena < lenb) ? -1 : 1;
        }
    }

    return cmp;
}

bool tsquery_requires_match

(

QueryItem *

curitem

)

Definition at line 1865 of file tsvector_op.c.

References check_stack_depth(), elog, ERROR, QueryOperator::left, OP_AND, OP_NOT, OP_OR, OP_PHRASE, QueryOperator::oper, QI_VAL, QueryItem::qoperator, tsquery_requires_match(), and QueryItem::type.

Referenced by gin_extract_tsquery(), and tsquery_requires_match().

{
    /* since this function recurses, it could be driven to stack overflow */
    check_stack_depth();

    if (curitem->type == QI_VAL)
        return true;

    switch (curitem->qoperator.oper)
    {
        case OP_NOT:

            /*
             * Assume there are no required matches underneath a NOT.  For
             * some cases with nested NOTs, we could prove there's a required
             * match, but it seems unlikely to be worth the trouble.
             */
            return false;

        case OP_PHRASE:

            /*
             * Treat OP_PHRASE as OP_AND here
             */
        case OP_AND:
            /* If either side requires a match, we're good */
            if (tsquery_requires_match(curitem + curitem->qoperator.left))
                return true;
            else
                return tsquery_requires_match(curitem + 1);

        case OP_OR:
            /* Both sides must require a match */
            if (tsquery_requires_match(curitem + curitem->qoperator.left))
                return tsquery_requires_match(curitem + 1);
            else
                return false;

        default:
            elog(ERROR, "unrecognized operator: %d", curitem->qoperator.oper);
    }

    /* not reachable, but keep compiler quiet */
    return false;
}

static int tsvector_bsearch ( const TSVector  tsv, char *  lexeme, int  lexeme_len  )

static

Definition at line 393 of file tsvector_op.c.

References ARRPTR, cmp(), TSVectorData::size, STRPTR, and tsCompareString().

Referenced by tsvector_delete_arr(), tsvector_delete_str(), and tsvector_setweight_by_filter().

{
    WordEntry  *arrin = ARRPTR(tsv);
    int         StopLow = 0,
                StopHigh = tsv->size,
                StopMiddle,
                cmp;

    while (StopLow < StopHigh)
    {
        StopMiddle = (StopLow + StopHigh) / 2;

        cmp = tsCompareString(lexeme, lexeme_len,
                              STRPTR(tsv) + arrin[StopMiddle].pos,
                              arrin[StopMiddle].len,
                              false);

        if (cmp < 0)
            StopHigh = StopMiddle;
        else if (cmp > 0)
            StopLow = StopMiddle + 1;
        else                    /* found it */
            return StopMiddle;
    }

    return -1;
}

Datum tsvector_concat

(

PG_FUNCTION_ARGS

)

Definition at line 933 of file tsvector_op.c.

References _POSVECPTR, add_pos(), ARRPTR, Assert, CALCDATASIZE, cmp(), compareEntry, ereport, errcode(), errmsg(), ERROR, WordEntry::haspos, i, WordEntry::len, MAXSTRPOS, memmove, palloc0(), PG_FREE_IF_COPY, PG_GETARG_TSVECTOR, PG_RETURN_POINTER, WordEntry::pos, POSDATALEN, POSDATAPTR, SET_VARSIZE, SHORTALIGN, TSVectorData::size, STRPTR, VARSIZE, and WEP_GETPOS.

{
    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_delete_arr

(

PG_FUNCTION_ARGS

)

Definition at line 580 of file tsvector_op.c.

References deconstruct_array(), ereport, errcode(), errmsg(), ERROR, i, palloc0(), pfree(), PG_FREE_IF_COPY, PG_GETARG_ARRAYTYPE_P, PG_GETARG_TSVECTOR, PG_RETURN_POINTER, TEXTOID, tsvector_bsearch(), tsvector_delete_by_indices(), VARDATA, VARHDRSZ, and VARSIZE.

{
    TSVector    tsin = PG_GETARG_TSVECTOR(0),
                tsout;
    ArrayType  *lexemes = PG_GETARG_ARRAYTYPE_P(1);
    int         i,
                nlex,
                skip_count,
               *skip_indices;
    Datum      *dlexemes;
    bool       *nulls;

    deconstruct_array(lexemes, TEXTOID, -1, false, 'i',
                      &dlexemes, &nulls, &nlex);

    /*
     * In typical use case array of lexemes to delete is relatively small. So
     * here we optimize things for that scenario: iterate through lexarr
     * performing binary search of each lexeme from lexarr in tsvector.
     */
    skip_indices = palloc0(nlex * sizeof(int));
    for (i = skip_count = 0; i < nlex; i++)
    {
        char       *lex;
        int         lex_len,
                    lex_pos;

        if (nulls[i])
            ereport(ERROR,
                    (errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED),
                     errmsg("lexeme array may not contain nulls")));

        lex = VARDATA(dlexemes[i]);
        lex_len = VARSIZE(dlexemes[i]) - VARHDRSZ;
        lex_pos = tsvector_bsearch(tsin, lex, lex_len);

        if (lex_pos >= 0)
            skip_indices[skip_count++] = lex_pos;
    }

    tsout = tsvector_delete_by_indices(tsin, skip_indices, skip_count);

    pfree(skip_indices);
    PG_FREE_IF_COPY(tsin, 0);
    PG_FREE_IF_COPY(lexemes, 1);

    PG_RETURN_POINTER(tsout);
}

static TSVector tsvector_delete_by_indices ( TSVector  tsv, int *  indices_to_delete, int  indices_count  )

static

Definition at line 459 of file tsvector_op.c.

References ARRPTR, Assert, CALCDATASIZE, compare_int(), WordEntry::haspos, i, WordEntry::len, palloc0(), POSDATALEN, qsort, SET_VARSIZE, SHORTALIGN, TSVectorData::size, STRPTR, and VARSIZE.

Referenced by tsvector_delete_arr(), and tsvector_delete_str().

{
    TSVector    tsout;
    WordEntry  *arrin = ARRPTR(tsv),
               *arrout;
    char       *data = STRPTR(tsv),
               *dataout;
    int         i,              /* index in arrin */
                j,              /* index in arrout */
                k,              /* index in indices_to_delete */
                curoff;         /* index in dataout area */

    /*
     * Sort the filter array to simplify membership checks below.  Also, get
     * rid of any duplicate entries, so that we can assume that indices_count
     * is exactly equal to the number of lexemes that will be removed.
     */
    if (indices_count > 1)
    {
        int         kp;

        qsort(indices_to_delete, indices_count, sizeof(int), compare_int);
        kp = 0;
        for (k = 1; k < indices_count; k++)
        {
            if (indices_to_delete[k] != indices_to_delete[kp])
                indices_to_delete[++kp] = indices_to_delete[k];
        }
        indices_count = ++kp;
    }

    /*
     * Here we overestimate tsout size, since we don't know how much space is
     * used by the deleted lexeme(s).  We will set exact size below.
     */
    tsout = (TSVector) palloc0(VARSIZE(tsv));

    /* This count must be correct because STRPTR(tsout) relies on it. */
    tsout->size = tsv->size - indices_count;

    /*
     * Copy tsv to tsout, skipping lexemes listed in indices_to_delete.
     */
    arrout = ARRPTR(tsout);
    dataout = STRPTR(tsout);
    curoff = 0;
    for (i = j = k = 0; i < tsv->size; i++)
    {
        /*
         * If current i is present in indices_to_delete, skip this lexeme.
         * Since indices_to_delete is already sorted, we only need to check
         * the current (k'th) entry.
         */
        if (k < indices_count && i == indices_to_delete[k])
        {
            k++;
            continue;
        }

        /* Copy lexeme and its positions and weights */
        memcpy(dataout + curoff, data + arrin[i].pos, arrin[i].len);
        arrout[j].haspos = arrin[i].haspos;
        arrout[j].len = arrin[i].len;
        arrout[j].pos = curoff;
        curoff += arrin[i].len;
        if (arrin[i].haspos)
        {
            int         len = POSDATALEN(tsv, arrin + i) * sizeof(WordEntryPos)
            + sizeof(uint16);

            curoff = SHORTALIGN(curoff);
            memcpy(dataout + curoff,
                   STRPTR(tsv) + SHORTALIGN(arrin[i].pos + arrin[i].len),
                   len);
            curoff += len;
        }

        j++;
    }

    /*
     * k should now be exactly equal to indices_count. If it isn't then the
     * caller provided us with indices outside of [0, tsv->size) range and
     * estimation of tsout's size is wrong.
     */
    Assert(k == indices_count);

    SET_VARSIZE(tsout, CALCDATASIZE(tsout->size, curoff));
    return tsout;
}

Datum tsvector_delete_str

(

PG_FUNCTION_ARGS

)

Definition at line 556 of file tsvector_op.c.

References StatEntry::lexeme, PG_FREE_IF_COPY, PG_GETARG_TEXT_PP, PG_GETARG_TSVECTOR, PG_RETURN_POINTER, tsvector_bsearch(), tsvector_delete_by_indices(), VARDATA_ANY, and VARSIZE_ANY_EXHDR.

{
    TSVector    tsin = PG_GETARG_TSVECTOR(0),
                tsout;
    text       *tlexeme = PG_GETARG_TEXT_PP(1);
    char       *lexeme = VARDATA_ANY(tlexeme);
    int         lexeme_len = VARSIZE_ANY_EXHDR(tlexeme),
                skip_index;

    if ((skip_index = tsvector_bsearch(tsin, lexeme, lexeme_len)) == -1)
        PG_RETURN_POINTER(tsin);

    tsout = tsvector_delete_by_indices(tsin, &skip_index, 1);

    PG_FREE_IF_COPY(tsin, 0);
    PG_FREE_IF_COPY(tlexeme, 1);
    PG_RETURN_POINTER(tsout);
}

Datum tsvector_filter

(

PG_FUNCTION_ARGS

)

Definition at line 826 of file tsvector_op.c.

References _POSVECPTR, ARRPTR, CALCDATASIZE, CHAROID, DatumGetChar, deconstruct_array(), ereport, errcode(), errmsg(), ERROR, i, WordEntry::len, memmove, WordEntryPosVector::npos, palloc0(), PG_FREE_IF_COPY, PG_GETARG_ARRAYTYPE_P, PG_GETARG_TSVECTOR, PG_RETURN_POINTER, WordEntryPosVector::pos, POSDATALEN, SET_VARSIZE, SHORTALIGN, TSVectorData::size, STRPTR, VARSIZE, weights, and WEP_GETWEIGHT.

{
    TSVector    tsin = PG_GETARG_TSVECTOR(0),
                tsout;
    ArrayType  *weights = PG_GETARG_ARRAYTYPE_P(1);
    WordEntry  *arrin = ARRPTR(tsin),
               *arrout;
    char       *datain = STRPTR(tsin),
               *dataout;
    Datum      *dweights;
    bool       *nulls;
    int         nweights;
    int         i,
                j;
    int         cur_pos = 0;
    char        mask = 0;

    deconstruct_array(weights, CHAROID, 1, true, 'c',
                      &dweights, &nulls, &nweights);

    for (i = 0; i < nweights; i++)
    {
        char        char_weight;

        if (nulls[i])
            ereport(ERROR,
                    (errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED),
                     errmsg("weight array may not contain nulls")));

        char_weight = DatumGetChar(dweights[i]);
        switch (char_weight)
        {
            case 'A':
            case 'a':
                mask = mask | 8;
                break;
            case 'B':
            case 'b':
                mask = mask | 4;
                break;
            case 'C':
            case 'c':
                mask = mask | 2;
                break;
            case 'D':
            case 'd':
                mask = mask | 1;
                break;
            default:
                ereport(ERROR,
                        (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                         errmsg("unrecognized weight: \"%c\"", char_weight)));
        }
    }

    tsout = (TSVector) palloc0(VARSIZE(tsin));
    tsout->size = tsin->size;
    arrout = ARRPTR(tsout);
    dataout = STRPTR(tsout);

    for (i = j = 0; i < tsin->size; i++)
    {
        WordEntryPosVector *posvin,
                   *posvout;
        int         npos = 0;
        int         k;

        if (!arrin[i].haspos)
            continue;

        posvin = _POSVECPTR(tsin, arrin + i);
        posvout = (WordEntryPosVector *)
            (dataout + SHORTALIGN(cur_pos + arrin[i].len));

        for (k = 0; k < posvin->npos; k++)
        {
            if (mask & (1 << WEP_GETWEIGHT(posvin->pos[k])))
                posvout->pos[npos++] = posvin->pos[k];
        }

        /* if no satisfactory positions found, skip lexeme */
        if (!npos)
            continue;

        arrout[j].haspos = true;
        arrout[j].len = arrin[i].len;
        arrout[j].pos = cur_pos;

        memcpy(dataout + cur_pos, datain + arrin[i].pos, arrin[i].len);
        posvout->npos = npos;
        cur_pos += SHORTALIGN(arrin[i].len);
        cur_pos += POSDATALEN(tsout, arrout + j) * sizeof(WordEntryPos) +
            sizeof(uint16);
        j++;
    }

    tsout->size = j;
    if (dataout != STRPTR(tsout))
        memmove(STRPTR(tsout), dataout, cur_pos);

    SET_VARSIZE(tsout, CALCDATASIZE(tsout->size, cur_pos));

    PG_FREE_IF_COPY(tsin, 0);
    PG_RETURN_POINTER(tsout);
}

Datum tsvector_length

(

PG_FUNCTION_ARGS

)

Definition at line 192 of file tsvector_op.c.

References PG_FREE_IF_COPY, PG_GETARG_TSVECTOR, PG_RETURN_INT32, and TSVectorData::size.

{
    TSVector    in = PG_GETARG_TSVECTOR(0);
    int32       ret = in->size;

    PG_FREE_IF_COPY(in, 0);
    PG_RETURN_INT32(ret);
}

Datum tsvector_setweight

(

PG_FUNCTION_ARGS

)

Definition at line 202 of file tsvector_op.c.

References ARRPTR, elog, ERROR, i, palloc(), PG_FREE_IF_COPY, PG_GETARG_CHAR, PG_GETARG_TSVECTOR, PG_RETURN_POINTER, POSDATALEN, POSDATAPTR, TSVectorData::size, VARSIZE, and WEP_SETWEIGHT.

{
    TSVector    in = PG_GETARG_TSVECTOR(0);
    char        cw = PG_GETARG_CHAR(1);
    TSVector    out;
    int         i,
                j;
    WordEntry  *entry;
    WordEntryPos *p;
    int         w = 0;

    switch (cw)
    {
        case 'A':
        case 'a':
            w = 3;
            break;
        case 'B':
        case 'b':
            w = 2;
            break;
        case 'C':
        case 'c':
            w = 1;
            break;
        case 'D':
        case 'd':
            w = 0;
            break;
        default:
            /* internal error */
            elog(ERROR, "unrecognized weight: %d", cw);
    }

    out = (TSVector) palloc(VARSIZE(in));
    memcpy(out, in, VARSIZE(in));
    entry = ARRPTR(out);
    i = out->size;
    while (i--)
    {
        if ((j = POSDATALEN(out, entry)) != 0)
        {
            p = POSDATAPTR(out, entry);
            while (j--)
            {
                WEP_SETWEIGHT(*p, w);
                p++;
            }
        }
        entry++;
    }

    PG_FREE_IF_COPY(in, 0);
    PG_RETURN_POINTER(out);
}

Datum tsvector_setweight_by_filter

(

PG_FUNCTION_ARGS

)

Definition at line 264 of file tsvector_op.c.

References ARRPTR, deconstruct_array(), elog, ereport, errcode(), errmsg(), ERROR, i, palloc(), PG_FREE_IF_COPY, PG_GETARG_ARRAYTYPE_P, PG_GETARG_CHAR, PG_GETARG_TSVECTOR, PG_RETURN_POINTER, POSDATALEN, POSDATAPTR, TEXTOID, tsvector_bsearch(), VARDATA, VARHDRSZ, VARSIZE, and WEP_SETWEIGHT.

{
    TSVector    tsin = PG_GETARG_TSVECTOR(0);
    char        char_weight = PG_GETARG_CHAR(1);
    ArrayType  *lexemes = PG_GETARG_ARRAYTYPE_P(2);

    TSVector    tsout;
    int         i,
                j,
                nlexemes,
                weight;
    WordEntry  *entry;
    Datum      *dlexemes;
    bool       *nulls;

    switch (char_weight)
    {
        case 'A':
        case 'a':
            weight = 3;
            break;
        case 'B':
        case 'b':
            weight = 2;
            break;
        case 'C':
        case 'c':
            weight = 1;
            break;
        case 'D':
        case 'd':
            weight = 0;
            break;
        default:
            /* internal error */
            elog(ERROR, "unrecognized weight: %c", char_weight);
    }

    tsout = (TSVector) palloc(VARSIZE(tsin));
    memcpy(tsout, tsin, VARSIZE(tsin));
    entry = ARRPTR(tsout);

    deconstruct_array(lexemes, TEXTOID, -1, false, 'i',
                      &dlexemes, &nulls, &nlexemes);

    /*
     * Assuming that lexemes array is significantly shorter than tsvector we
     * can iterate through lexemes performing binary search of each lexeme
     * from lexemes in tsvector.
     */
    for (i = 0; i < nlexemes; i++)
    {
        char       *lex;
        int         lex_len,
                    lex_pos;

        if (nulls[i])
            ereport(ERROR,
                    (errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED),
                     errmsg("lexeme array may not contain nulls")));

        lex = VARDATA(dlexemes[i]);
        lex_len = VARSIZE(dlexemes[i]) - VARHDRSZ;
        lex_pos = tsvector_bsearch(tsout, lex, lex_len);

        if (lex_pos >= 0 && (j = POSDATALEN(tsout, entry + lex_pos)) != 0)
        {
            WordEntryPos *p = POSDATAPTR(tsout, entry + lex_pos);

            while (j--)
            {
                WEP_SETWEIGHT(*p, weight);
                p++;
            }
        }
    }

    PG_FREE_IF_COPY(tsin, 0);
    PG_FREE_IF_COPY(lexemes, 2);

    PG_RETURN_POINTER(tsout);
}

Datum tsvector_strip

(

PG_FUNCTION_ARGS

)

Definition at line 159 of file tsvector_op.c.

References ARRPTR, CALCDATASIZE, cur, i, WordEntry::len, palloc0(), PG_FREE_IF_COPY, PG_GETARG_TSVECTOR, PG_RETURN_POINTER, SET_VARSIZE, TSVectorData::size, and STRPTR.

{
    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);
}

Datum tsvector_to_array

(

PG_FUNCTION_ARGS

)

Definition at line 728 of file tsvector_op.c.

References ARRPTR, construct_array(), cstring_to_text_with_len(), i, palloc(), pfree(), PG_FREE_IF_COPY, PG_GETARG_TSVECTOR, PG_RETURN_POINTER, PointerGetDatum, TSVectorData::size, STRPTR, and TEXTOID.

{
    TSVector    tsin = PG_GETARG_TSVECTOR(0);
    WordEntry  *arrin = ARRPTR(tsin);
    Datum      *elements;
    int         i;
    ArrayType  *array;

    elements = palloc(tsin->size * sizeof(Datum));

    for (i = 0; i < tsin->size; i++)
    {
        elements[i] = PointerGetDatum(
                                      cstring_to_text_with_len(STRPTR(tsin) + arrin[i].pos, arrin[i].len)
            );
    }

    array = construct_array(elements, tsin->size, TEXTOID, -1, false, 'i');

    pfree(elements);
    PG_FREE_IF_COPY(tsin, 0);
    PG_RETURN_POINTER(array);
}

Datum tsvector_unnest

(

PG_FUNCTION_ARGS

)

Definition at line 636 of file tsvector_op.c.

References _POSVECPTR, ARRPTR, BlessTupleDesc(), FuncCallContext::call_cntr, construct_array(), CreateTemplateTupleDesc(), cstring_to_text_with_len(), heap_form_tuple(), HeapTupleGetDatum, i, Int16GetDatum, INT2ARRAYOID, INT2OID, MemoryContextSwitchTo(), FuncCallContext::multi_call_memory_ctx, WordEntryPosVector::npos, palloc(), pfree(), PG_GETARG_TSVECTOR_COPY, PointerGetDatum, WordEntryPosVector::pos, TSVectorData::size, SRF_FIRSTCALL_INIT, SRF_IS_FIRSTCALL, SRF_PERCALL_SETUP, SRF_RETURN_DONE, SRF_RETURN_NEXT, STRPTR, TEXTARRAYOID, TEXTOID, FuncCallContext::tuple_desc, TupleDescInitEntry(), FuncCallContext::user_fctx, values, weights, WEP_GETPOS, and WEP_GETWEIGHT.

{
    FuncCallContext *funcctx;
    TSVector    tsin;

    if (SRF_IS_FIRSTCALL())
    {
        MemoryContext oldcontext;
        TupleDesc   tupdesc;

        funcctx = SRF_FIRSTCALL_INIT();
        oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);

        tupdesc = CreateTemplateTupleDesc(3, false);
        TupleDescInitEntry(tupdesc, (AttrNumber) 1, "lexeme",
                           TEXTOID, -1, 0);
        TupleDescInitEntry(tupdesc, (AttrNumber) 2, "positions",
                           INT2ARRAYOID, -1, 0);
        TupleDescInitEntry(tupdesc, (AttrNumber) 3, "weights",
                           TEXTARRAYOID, -1, 0);
        funcctx->tuple_desc = BlessTupleDesc(tupdesc);

        funcctx->user_fctx = PG_GETARG_TSVECTOR_COPY(0);

        MemoryContextSwitchTo(oldcontext);
    }

    funcctx = SRF_PERCALL_SETUP();
    tsin = (TSVector) funcctx->user_fctx;

    if (funcctx->call_cntr < tsin->size)
    {
        WordEntry  *arrin = ARRPTR(tsin);
        char       *data = STRPTR(tsin);
        HeapTuple   tuple;
        int         j,
                    i = funcctx->call_cntr;
        bool        nulls[] = {false, false, false};
        Datum       values[3];

        values[0] = PointerGetDatum(
                                    cstring_to_text_with_len(data + arrin[i].pos, arrin[i].len)
            );

        if (arrin[i].haspos)
        {
            WordEntryPosVector *posv;
            Datum      *positions;
            Datum      *weights;
            char        weight;

            /*
             * Internally tsvector stores position and weight in the same
             * uint16 (2 bits for weight, 14 for position). Here we extract
             * that in two separate arrays.
             */
            posv = _POSVECPTR(tsin, arrin + i);
            positions = palloc(posv->npos * sizeof(Datum));
            weights = palloc(posv->npos * sizeof(Datum));
            for (j = 0; j < posv->npos; j++)
            {
                positions[j] = Int16GetDatum(WEP_GETPOS(posv->pos[j]));
                weight = 'D' - WEP_GETWEIGHT(posv->pos[j]);
                weights[j] = PointerGetDatum(
                                             cstring_to_text_with_len(&weight, 1)
                    );
            }

            values[1] = PointerGetDatum(
                                        construct_array(positions, posv->npos, INT2OID, 2, true, 's'));
            values[2] = PointerGetDatum(
                                        construct_array(weights, posv->npos, TEXTOID, -1, false, 'i'));
        }
        else
        {
            nulls[1] = nulls[2] = true;
        }

        tuple = heap_form_tuple(funcctx->tuple_desc, values, nulls);
        SRF_RETURN_NEXT(funcctx, HeapTupleGetDatum(tuple));
    }
    else
    {
        pfree(tsin);
        SRF_RETURN_DONE(funcctx);
    }
}

static Datum tsvector_update_trigger ( PG_FUNCTION_ARGS  , bool  config_column  )

static

Definition at line 2454 of file tsvector_op.c.

References CALLED_AS_TRIGGER, ParsedText::curwords, DatumGetObjectId, DatumGetPointer, DatumGetTextPP, elog, ereport, errcode(), errmsg(), ERROR, get_ts_config_oid(), heap_modify_tuple_by_cols(), i, IsBinaryCoercible(), ParsedText::lenwords, list_length(), make_tsvector(), numattr, palloc(), parsetext(), pfree(), PointerGetDatum, ParsedText::pos, RelationData::rd_att, REGCONFIGOID, SPI_ERROR_NOATTRIBUTE, SPI_fnumber(), SPI_getbinval(), SPI_gettypeid(), stringToQualifiedNameList(), TEXTOID, TriggerData::tg_event, TriggerData::tg_newtuple, TriggerData::tg_relation, TriggerData::tg_trigger, TriggerData::tg_trigtuple, Trigger::tgargs, Trigger::tgnargs, TRIGGER_FIRED_BEFORE, TRIGGER_FIRED_BY_INSERT, TRIGGER_FIRED_BY_UPDATE, TRIGGER_FIRED_FOR_ROW, TSVectorGetDatum, TSVECTOROID, VARDATA_ANY, VARSIZE_ANY_EXHDR, and ParsedText::words.

Referenced by tsvector_update_trigger_bycolumn(), and tsvector_update_trigger_byid().

{
    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])));
    /* This will effectively reject system columns, so no separate test: */
    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 = DatumGetTextPP(datum);

        parsetext(cfgId, &prs, VARDATA_ANY(txt), VARSIZE_ANY_EXHDR(txt));

        if (txt != (text *) DatumGetPointer(datum))
            pfree(txt);
    }

    /* make tsvector value */
    datum = TSVectorGetDatum(make_tsvector(&prs));
    isnull = false;

    /* and insert it into tuple */
    rettuple = heap_modify_tuple_by_cols(rettuple, rel->rd_att,
                                         1, &tsvector_attr_num,
                                         &datum, &isnull);

    pfree(DatumGetPointer(datum));

    return PointerGetDatum(rettuple);
}

Datum tsvector_update_trigger_bycolumn

(

PG_FUNCTION_ARGS

)

Definition at line 2448 of file tsvector_op.c.

References tsvector_update_trigger().

{
    return tsvector_update_trigger(fcinfo, true);
}

Datum tsvector_update_trigger_byid

(

PG_FUNCTION_ARGS

)

Definition at line 2442 of file tsvector_op.c.

References tsvector_update_trigger().

{
    return tsvector_update_trigger(fcinfo, false);
}

TSVECTORCMPFUNC

(

lt

)

TSVECTORCMPFUNC	(	le	,
		<=	,
		BOOL
	)

TSVECTORCMPFUNC	(	eq	,
		BOOL
	)

TSVECTORCMPFUNC	(	ge	,
		>=	,
		BOOL
	)

TSVECTORCMPFUNC	(	gt	,
		BOOL
	)

TSVECTORCMPFUNC	(	ne	,
		!	,
		BOOL
	)

TSVECTORCMPFUNC	(	cmp	,
		+	,
		INT32
	)

static int uniqueLongPos ( WordEntryPos *  pos, int  npos  )

static

Definition at line 1281 of file tsvector_op.c.

References compareWordEntryPos(), qsort, result, and WEP_GETPOS.

Referenced by checkcondition_str().

{
    WordEntryPos *pos_iter,
               *result;

    if (npos <= 1)
        return npos;

    qsort((void *) pos, npos, sizeof(WordEntryPos), compareWordEntryPos);

    result = pos;
    pos_iter = pos + 1;
    while (pos_iter < pos + npos)
    {
        if (WEP_GETPOS(*pos_iter) != WEP_GETPOS(*result))
        {
            result++;
            *result = WEP_GETPOS(*pos_iter);
        }

        pos_iter++;
    }

    return result + 1 - pos;
}

static StatEntry* walkStatEntryTree ( TSVectorStat *  stat )

static

Definition at line 2204 of file tsvector_op.c.

References Assert, StatEntry::left, TSVectorStat::maxdepth, StatEntry::ndoc, StatEntry::right, TSVectorStat::stack, and TSVectorStat::stackpos.

Referenced by ts_process_call().

{
    StatEntry  *node = stat->stack[stat->stackpos];

    if (node == NULL)
        return NULL;

    if (node->ndoc != 0)
    {
        /* return entry itself: we already was at left sublink */
        return node;
    }
    else if (node->right && node->right != stat->stack[stat->stackpos + 1])
    {
        /* go on right sublink */
        stat->stackpos++;
        node = node->right;

        /* find most-left value */
        for (;;)
        {
            stat->stack[stat->stackpos] = node;
            if (node->left)
            {
                stat->stackpos++;
                node = node->left;
            }
            else
                break;
        }
        Assert(stat->stackpos <= stat->maxdepth);
    }
    else
    {
        /* we already return all left subtree, itself and  right subtree */
        if (stat->stackpos == 0)
            return NULL;

        stat->stackpos--;
        return walkStatEntryTree(stat);
    }

    return node;
}