selfuncs.h File Reference

#include "access/htup.h"
#include "fmgr.h"
#include "nodes/pathnodes.h"

Include dependency graph for selfuncs.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Data Structures
struct	EstimationInfo
struct	VariableStatData
struct	GenericCosts

Macros
#define	DEFAULT_EQ_SEL   0.005
#define	DEFAULT_INEQ_SEL   0.3333333333333333
#define	DEFAULT_RANGE_INEQ_SEL   0.005
#define	DEFAULT_MULTIRANGE_INEQ_SEL   0.005
#define	DEFAULT_MATCH_SEL   0.005
#define	DEFAULT_MATCHING_SEL   0.010
#define	DEFAULT_NUM_DISTINCT   200
#define	DEFAULT_UNK_SEL   0.005
#define	DEFAULT_NOT_UNK_SEL   (1.0 - DEFAULT_UNK_SEL)
#define	CLAMP_PROBABILITY(p)
#define	SELFLAG_USED_DEFAULT
#define	ReleaseVariableStats(vardata)

Typedefs
typedef struct EstimationInfo	EstimationInfo
typedef struct VariableStatData	VariableStatData
typedef bool(*	get_relation_stats_hook_type) (PlannerInfo *root, RangeTblEntry *rte, AttrNumber attnum, VariableStatData *vardata)
typedef bool(*	get_index_stats_hook_type) (PlannerInfo *root, Oid indexOid, AttrNumber indexattnum, VariableStatData *vardata)

Functions
void	examine_variable (PlannerInfo *root, Node *node, int varRelid, VariableStatData *vardata)
bool	all_rows_selectable (PlannerInfo *root, Index varno, Bitmapset *varattnos)
bool	statistic_proc_security_check (VariableStatData *vardata, Oid func_oid)
bool	get_restriction_variable (PlannerInfo *root, List *args, int varRelid, VariableStatData *vardata, Node **other, bool *varonleft)
void	get_join_variables (PlannerInfo *root, List *args, SpecialJoinInfo *sjinfo, VariableStatData *vardata1, VariableStatData *vardata2, bool *join_is_reversed)
double	get_variable_numdistinct (VariableStatData *vardata, bool *isdefault)
double	mcv_selectivity (VariableStatData *vardata, FmgrInfo *opproc, Oid collation, Datum constval, bool varonleft, double *sumcommonp)
double	histogram_selectivity (VariableStatData *vardata, FmgrInfo *opproc, Oid collation, Datum constval, bool varonleft, int min_hist_size, int n_skip, int *hist_size)
double	generic_restriction_selectivity (PlannerInfo *root, Oid oproid, Oid collation, List *args, int varRelid, double default_selectivity)
double	ineq_histogram_selectivity (PlannerInfo *root, VariableStatData *vardata, Oid opoid, FmgrInfo *opproc, bool isgt, bool iseq, Oid collation, Datum constval, Oid consttype)
double	var_eq_const (VariableStatData *vardata, Oid oproid, Oid collation, Datum constval, bool constisnull, bool varonleft, bool negate)
double	var_eq_non_const (VariableStatData *vardata, Oid oproid, Oid collation, Node *other, bool varonleft, bool negate)
Selectivity	boolvarsel (PlannerInfo *root, Node *arg, int varRelid)
Selectivity	booltestsel (PlannerInfo *root, BoolTestType booltesttype, Node *arg, int varRelid, JoinType jointype, SpecialJoinInfo *sjinfo)
Selectivity	nulltestsel (PlannerInfo *root, NullTestType nulltesttype, Node *arg, int varRelid, JoinType jointype, SpecialJoinInfo *sjinfo)
Selectivity	scalararraysel (PlannerInfo *root, ScalarArrayOpExpr *clause, bool is_join_clause, int varRelid, JoinType jointype, SpecialJoinInfo *sjinfo)
double	estimate_array_length (PlannerInfo *root, Node *arrayexpr)
Selectivity	rowcomparesel (PlannerInfo *root, RowCompareExpr *clause, int varRelid, JoinType jointype, SpecialJoinInfo *sjinfo)
void	mergejoinscansel (PlannerInfo *root, Node *clause, Oid opfamily, CompareType cmptype, bool nulls_first, Selectivity *leftstart, Selectivity *leftend, Selectivity *rightstart, Selectivity *rightend)
double	estimate_num_groups (PlannerInfo *root, List *groupExprs, double input_rows, List **pgset, EstimationInfo *estinfo)
List *	estimate_multivariate_bucketsize (PlannerInfo *root, RelOptInfo *inner, List *hashclauses, Selectivity *innerbucketsize)
void	estimate_hash_bucket_stats (PlannerInfo *root, Node *hashkey, double nbuckets, Selectivity *mcv_freq, Selectivity *bucketsize_frac)
double	estimate_hashagg_tablesize (PlannerInfo *root, Path *path, const AggClauseCosts *agg_costs, double dNumGroups)
List *	get_quals_from_indexclauses (List *indexclauses)
Cost	index_other_operands_eval_cost (PlannerInfo *root, List *indexquals)
List *	add_predicate_to_index_quals (IndexOptInfo *index, List *indexQuals)
void	genericcostestimate (PlannerInfo *root, IndexPath *path, double loop_count, GenericCosts *costs)
Selectivity	scalararraysel_containment (PlannerInfo *root, Node *leftop, Node *rightop, Oid elemtype, bool isEquality, bool useOr, int varRelid)

Variables
PGDLLIMPORT get_relation_stats_hook_type	get_relation_stats_hook
PGDLLIMPORT get_index_stats_hook_type	get_index_stats_hook

Macro Definition Documentation

CLAMP_PROBABILITY

#define CLAMP_PROBABILITY

(

p

)

Value:

    do { \
        if (p < 0.0) \
            p = 0.0; \
        else if (p > 1.0) \
            p = 1.0; \
    } while (0)

Definition at line 63 of file selfuncs.h.

DEFAULT_EQ_SEL

#define DEFAULT_EQ_SEL   0.005

Definition at line 34 of file selfuncs.h.

DEFAULT_INEQ_SEL

#define DEFAULT_INEQ_SEL   0.3333333333333333

Definition at line 37 of file selfuncs.h.

DEFAULT_MATCH_SEL

#define DEFAULT_MATCH_SEL   0.005

Definition at line 46 of file selfuncs.h.

DEFAULT_MATCHING_SEL

#define DEFAULT_MATCHING_SEL   0.010

Definition at line 49 of file selfuncs.h.

DEFAULT_MULTIRANGE_INEQ_SEL

#define DEFAULT_MULTIRANGE_INEQ_SEL   0.005

Definition at line 43 of file selfuncs.h.

DEFAULT_NOT_UNK_SEL

#define DEFAULT_NOT_UNK_SEL   (1.0 - DEFAULT_UNK_SEL)

Definition at line 56 of file selfuncs.h.

DEFAULT_NUM_DISTINCT

#define DEFAULT_NUM_DISTINCT   200

Definition at line 52 of file selfuncs.h.

DEFAULT_RANGE_INEQ_SEL

#define DEFAULT_RANGE_INEQ_SEL   0.005

Definition at line 40 of file selfuncs.h.

DEFAULT_UNK_SEL

#define DEFAULT_UNK_SEL   0.005

Definition at line 55 of file selfuncs.h.

ReleaseVariableStats

#define ReleaseVariableStats

(

vardata

)

Value:

    do { \
        if (HeapTupleIsValid((vardata).statsTuple)) \
            (vardata).freefunc((vardata).statsTuple); \
    } while(0)

Definition at line 101 of file selfuncs.h.

SELFLAG_USED_DEFAULT

#define SELFLAG_USED_DEFAULT

Value:

                                                 (1 << 0)   /* Estimation fell back on one
                                                 * of the DEFAULTs as defined
                                                 * above. */

Definition at line 76 of file selfuncs.h.

Typedef Documentation

EstimationInfo

typedef struct EstimationInfo EstimationInfo

get_index_stats_hook_type

typedef bool(* get_index_stats_hook_type) (PlannerInfo *root, Oid indexOid, AttrNumber indexattnum, VariableStatData *vardata)

Definition at line 145 of file selfuncs.h.

get_relation_stats_hook_type

typedef bool(* get_relation_stats_hook_type) (PlannerInfo *root, RangeTblEntry *rte, AttrNumber attnum, VariableStatData *vardata)

Definition at line 140 of file selfuncs.h.

VariableStatData

typedef struct VariableStatData VariableStatData

Function Documentation

add_predicate_to_index_quals()

List * add_predicate_to_index_quals	(	IndexOptInfo *	index,
		List *	indexQuals
	)

Definition at line 7618 of file selfuncs.c.

{
    List       *predExtraQuals = NIL;
    ListCell   *lc;
 
    if (index->indpred == NIL)
        return indexQuals;
 
    foreach(lc, index->indpred)
    {
        Node       *predQual = (Node *) lfirst(lc);
        List       *oneQual = list_make1(predQual);
 
        if (!predicate_implied_by(oneQual, indexQuals, false))
            predExtraQuals = list_concat(predExtraQuals, oneQual);
    }
    return list_concat(predExtraQuals, indexQuals);
}

References lfirst, list_concat(), list_make1, NIL, and predicate_implied_by().

Referenced by btcostestimate(), genericcostestimate(), and gincostestimate().

all_rows_selectable()

bool all_rows_selectable	(	PlannerInfo *	root,
		Index	varno,
		Bitmapset *	varattnos
	)

Definition at line 6304 of file selfuncs.c.

{
    RelOptInfo *rel = find_base_rel_noerr(root, varno);
    RangeTblEntry *rte = planner_rt_fetch(varno, root);
    Oid         userid;
    int         varattno;
 
    Assert(rte->rtekind == RTE_RELATION);
 
    /*
     * Determine the user ID to use for privilege checks (either the current
     * user or the view owner, if we're accessing the table via a view).
     *
     * Normally the relation will have an associated RelOptInfo from which we
     * can find the userid, but it might not if it's a RETURNING Var for an
     * INSERT target relation.  In that case use the RTEPermissionInfo
     * associated with the RTE.
     *
     * If we navigate up to a parent relation, we keep using the same userid,
     * since it's the same in all relations of a given inheritance tree.
     */
    if (rel)
        userid = rel->userid;
    else
    {
        RTEPermissionInfo *perminfo;
 
        perminfo = getRTEPermissionInfo(root->parse->rteperminfos, rte);
        userid = perminfo->checkAsUser;
    }
    if (!OidIsValid(userid))
        userid = GetUserId();
 
    /*
     * Permissions and securityQuals must be checked on the table actually
     * mentioned in the query, so if this is an inheritance child, navigate up
     * to the inheritance root parent.  If the user can read the whole table
     * or the required columns there, then they can read from the child table
     * too.  For per-column checks, we must find out which of the root
     * parent's attributes the child relation's attributes correspond to.
     */
    if (root->append_rel_array != NULL)
    {
        AppendRelInfo *appinfo;
 
        appinfo = root->append_rel_array[varno];
 
        /*
         * Partitions are mapped to their immediate parent, not the root
         * parent, so must be ready to walk up multiple AppendRelInfos.  But
         * stop if we hit a parent that is not RTE_RELATION --- that's a
         * flattened UNION ALL subquery, not an inheritance parent.
         */
        while (appinfo &&
               planner_rt_fetch(appinfo->parent_relid,
                                root)->rtekind == RTE_RELATION)
        {
            Bitmapset  *parent_varattnos = NULL;
 
            /*
             * For each child attribute, find the corresponding parent
             * attribute.  In rare cases, the attribute may be local to the
             * child table, in which case, we've got to live with having no
             * access to this column.
             */
            varattno = -1;
            while ((varattno = bms_next_member(varattnos, varattno)) >= 0)
            {
                AttrNumber  attno;
                AttrNumber  parent_attno;
 
                attno = varattno + FirstLowInvalidHeapAttributeNumber;
 
                if (attno == InvalidAttrNumber)
                {
                    /*
                     * Whole-row reference, so must map each column of the
                     * child to the parent table.
                     */
                    for (attno = 1; attno <= appinfo->num_child_cols; attno++)
                    {
                        parent_attno = appinfo->parent_colnos[attno - 1];
                        if (parent_attno == 0)
                            return false;   /* attr is local to child */
                        parent_varattnos =
                            bms_add_member(parent_varattnos,
                                           parent_attno - FirstLowInvalidHeapAttributeNumber);
                    }
                }
                else
                {
                    if (attno < 0)
                    {
                        /* System attnos are the same in all tables */
                        parent_attno = attno;
                    }
                    else
                    {
                        if (attno > appinfo->num_child_cols)
                            return false;   /* safety check */
                        parent_attno = appinfo->parent_colnos[attno - 1];
                        if (parent_attno == 0)
                            return false;   /* attr is local to child */
                    }
                    parent_varattnos =
                        bms_add_member(parent_varattnos,
                                       parent_attno - FirstLowInvalidHeapAttributeNumber);
                }
            }
 
            /* If the parent is itself a child, continue up */
            varno = appinfo->parent_relid;
            varattnos = parent_varattnos;
            appinfo = root->append_rel_array[varno];
        }
 
        /* Perform the access check on this parent rel */
        rte = planner_rt_fetch(varno, root);
        Assert(rte->rtekind == RTE_RELATION);
    }
 
    /*
     * For all rows to be accessible, there must be no securityQuals from
     * security barrier views or RLS policies.
     */
    if (rte->securityQuals != NIL)
        return false;
 
    /*
     * Test for table-level SELECT privilege.
     *
     * If varattnos is non-NULL, this is sufficient to give access to all
     * requested attributes, even for a child table, since we have verified
     * that all required child columns have matching parent columns.
     *
     * If varattnos is NULL (whole-table access requested), this doesn't
     * necessarily guarantee that the user can read all columns of a child
     * table, but we allow it anyway (see comments in examine_variable()) and
     * don't bother checking any column privileges.
     */
    if (pg_class_aclcheck(rte->relid, userid, ACL_SELECT) == ACLCHECK_OK)
        return true;
 
    if (varattnos == NULL)
        return false;           /* whole-table access requested */
 
    /*
     * Don't have table-level SELECT privilege, so check per-column
     * privileges.
     */
    varattno = -1;
    while ((varattno = bms_next_member(varattnos, varattno)) >= 0)
    {
        AttrNumber  attno = varattno + FirstLowInvalidHeapAttributeNumber;
 
        if (attno == InvalidAttrNumber)
        {
            /* Whole-row reference, so must have access to all columns */
            if (pg_attribute_aclcheck_all(rte->relid, userid, ACL_SELECT,
                                          ACLMASK_ALL) != ACLCHECK_OK)
                return false;
        }
        else
        {
            if (pg_attribute_aclcheck(rte->relid, attno, userid,
                                      ACL_SELECT) != ACLCHECK_OK)
                return false;
        }
    }
 
    /* If we reach here, have all required column privileges */
    return true;
}

References ACL_SELECT, ACLCHECK_OK, ACLMASK_ALL, Assert(), bms_add_member(), bms_next_member(), RTEPermissionInfo::checkAsUser, find_base_rel_noerr(), FirstLowInvalidHeapAttributeNumber, getRTEPermissionInfo(), GetUserId(), InvalidAttrNumber, NIL, AppendRelInfo::num_child_cols, OidIsValid, AppendRelInfo::parent_relid, pg_attribute_aclcheck(), pg_attribute_aclcheck_all(), pg_class_aclcheck(), planner_rt_fetch, root, RTE_RELATION, RangeTblEntry::rtekind, and RelOptInfo::userid.

Referenced by examine_simple_variable(), examine_variable(), and statext_is_compatible_clause().

booltestsel()

Selectivity booltestsel	(	PlannerInfo *	root,
		BoolTestType	booltesttype,
		Node *	arg,
		int	varRelid,
		JoinType	jointype,
		SpecialJoinInfo *	sjinfo
	)

Definition at line 1624 of file selfuncs.c.

{
    VariableStatData vardata;
    double      selec;
 
    examine_variable(root, arg, varRelid, &vardata);
 
    if (HeapTupleIsValid(vardata.statsTuple))
    {
        Form_pg_statistic stats;
        double      freq_null;
        AttStatsSlot sslot;
 
        stats = (Form_pg_statistic) GETSTRUCT(vardata.statsTuple);
        freq_null = stats->stanullfrac;
 
        if (get_attstatsslot(&sslot, vardata.statsTuple,
                             STATISTIC_KIND_MCV, InvalidOid,
                             ATTSTATSSLOT_VALUES | ATTSTATSSLOT_NUMBERS)
            && sslot.nnumbers > 0)
        {
            double      freq_true;
            double      freq_false;
 
            /*
             * Get first MCV frequency and derive frequency for true.
             */
            if (DatumGetBool(sslot.values[0]))
                freq_true = sslot.numbers[0];
            else
                freq_true = 1.0 - sslot.numbers[0] - freq_null;
 
            /*
             * Next derive frequency for false. Then use these as appropriate
             * to derive frequency for each case.
             */
            freq_false = 1.0 - freq_true - freq_null;
 
            switch (booltesttype)
            {
                case IS_UNKNOWN:
                    /* select only NULL values */
                    selec = freq_null;
                    break;
                case IS_NOT_UNKNOWN:
                    /* select non-NULL values */
                    selec = 1.0 - freq_null;
                    break;
                case IS_TRUE:
                    /* select only TRUE values */
                    selec = freq_true;
                    break;
                case IS_NOT_TRUE:
                    /* select non-TRUE values */
                    selec = 1.0 - freq_true;
                    break;
                case IS_FALSE:
                    /* select only FALSE values */
                    selec = freq_false;
                    break;
                case IS_NOT_FALSE:
                    /* select non-FALSE values */
                    selec = 1.0 - freq_false;
                    break;
                default:
                    elog(ERROR, "unrecognized booltesttype: %d",
                         (int) booltesttype);
                    selec = 0.0;    /* Keep compiler quiet */
                    break;
            }
 
            free_attstatsslot(&sslot);
        }
        else
        {
            /*
             * No most-common-value info available. Still have null fraction
             * information, so use it for IS [NOT] UNKNOWN. Otherwise adjust
             * for null fraction and assume a 50-50 split of TRUE and FALSE.
             */
            switch (booltesttype)
            {
                case IS_UNKNOWN:
                    /* select only NULL values */
                    selec = freq_null;
                    break;
                case IS_NOT_UNKNOWN:
                    /* select non-NULL values */
                    selec = 1.0 - freq_null;
                    break;
                case IS_TRUE:
                case IS_FALSE:
                    /* Assume we select half of the non-NULL values */
                    selec = (1.0 - freq_null) / 2.0;
                    break;
                case IS_NOT_TRUE:
                case IS_NOT_FALSE:
                    /* Assume we select NULLs plus half of the non-NULLs */
                    /* equiv. to freq_null + (1.0 - freq_null) / 2.0 */
                    selec = (freq_null + 1.0) / 2.0;
                    break;
                default:
                    elog(ERROR, "unrecognized booltesttype: %d",
                         (int) booltesttype);
                    selec = 0.0;    /* Keep compiler quiet */
                    break;
            }
        }
    }
    else
    {
        /*
         * If we can't get variable statistics for the argument, perhaps
         * clause_selectivity can do something with it.  We ignore the
         * possibility of a NULL value when using clause_selectivity, and just
         * assume the value is either TRUE or FALSE.
         */
        switch (booltesttype)
        {
            case IS_UNKNOWN:
                selec = DEFAULT_UNK_SEL;
                break;
            case IS_NOT_UNKNOWN:
                selec = DEFAULT_NOT_UNK_SEL;
                break;
            case IS_TRUE:
            case IS_NOT_FALSE:
                selec = (double) clause_selectivity(root, arg,
                                                    varRelid,
                                                    jointype, sjinfo);
                break;
            case IS_FALSE:
            case IS_NOT_TRUE:
                selec = 1.0 - (double) clause_selectivity(root, arg,
                                                          varRelid,
                                                          jointype, sjinfo);
                break;
            default:
                elog(ERROR, "unrecognized booltesttype: %d",
                     (int) booltesttype);
                selec = 0.0;    /* Keep compiler quiet */
                break;
        }
    }
 
    ReleaseVariableStats(vardata);
 
    /* result should be in range, but make sure... */
    CLAMP_PROBABILITY(selec);
 
    return (Selectivity) selec;
}

References arg, ATTSTATSSLOT_NUMBERS, ATTSTATSSLOT_VALUES, CLAMP_PROBABILITY, clause_selectivity(), DatumGetBool(), DEFAULT_NOT_UNK_SEL, DEFAULT_UNK_SEL, elog, ERROR, examine_variable(), free_attstatsslot(), get_attstatsslot(), GETSTRUCT(), HeapTupleIsValid, InvalidOid, IS_FALSE, IS_NOT_FALSE, IS_NOT_TRUE, IS_NOT_UNKNOWN, IS_TRUE, IS_UNKNOWN, AttStatsSlot::nnumbers, AttStatsSlot::numbers, ReleaseVariableStats, root, VariableStatData::statsTuple, and AttStatsSlot::values.

Referenced by clause_selectivity_ext().

boolvarsel()

Selectivity boolvarsel	(	PlannerInfo *	root,
		Node *	arg,
		int	varRelid
	)

Definition at line 1585 of file selfuncs.c.

{
    VariableStatData vardata;
    double      selec;
 
    examine_variable(root, arg, varRelid, &vardata);
    if (HeapTupleIsValid(vardata.statsTuple))
    {
        /*
         * A boolean variable V is equivalent to the clause V = 't', so we
         * compute the selectivity as if that is what we have.
         */
        selec = var_eq_const(&vardata, BooleanEqualOperator, InvalidOid,
                             BoolGetDatum(true), false, true, false);
    }
    else if (is_funcclause(arg))
    {
        /*
         * If we have no stats and it's a function call, estimate 0.3333333.
         * This seems a pretty unprincipled choice, but Postgres has been
         * using that estimate for function calls since 1992.  The hoariness
         * of this behavior suggests that we should not be in too much hurry
         * to use another value.
         */
        selec = 0.3333333;
    }
    else
    {
        /* Otherwise, the default estimate is 0.5 */
        selec = 0.5;
    }
    ReleaseVariableStats(vardata);
    return selec;
}

References arg, BoolGetDatum(), examine_variable(), HeapTupleIsValid, InvalidOid, is_funcclause(), ReleaseVariableStats, root, VariableStatData::statsTuple, and var_eq_const().

Referenced by clause_selectivity_ext().

estimate_array_length()

double estimate_array_length	(	PlannerInfo *	root,
		Node *	arrayexpr
	)

Definition at line 2223 of file selfuncs.c.

{
    /* look through any binary-compatible relabeling of arrayexpr */
    arrayexpr = strip_array_coercion(arrayexpr);
 
    if (arrayexpr && IsA(arrayexpr, Const))
    {
        Datum       arraydatum = ((Const *) arrayexpr)->constvalue;
        bool        arrayisnull = ((Const *) arrayexpr)->constisnull;
        ArrayType  *arrayval;
 
        if (arrayisnull)
            return 0;
        arrayval = DatumGetArrayTypeP(arraydatum);
        return ArrayGetNItems(ARR_NDIM(arrayval), ARR_DIMS(arrayval));
    }
    else if (arrayexpr && IsA(arrayexpr, ArrayExpr) &&
             !((ArrayExpr *) arrayexpr)->multidims)
    {
        return list_length(((ArrayExpr *) arrayexpr)->elements);
    }
    else if (arrayexpr && root)
    {
        /* See if we can find any statistics about it */
        VariableStatData vardata;
        AttStatsSlot sslot;
        double      nelem = 0;
 
        examine_variable(root, arrayexpr, 0, &vardata);
        if (HeapTupleIsValid(vardata.statsTuple))
        {
            /*
             * Found stats, so use the average element count, which is stored
             * in the last stanumbers element of the DECHIST statistics.
             * Actually that is the average count of *distinct* elements;
             * perhaps we should scale it up somewhat?
             */
            if (get_attstatsslot(&sslot, vardata.statsTuple,
                                 STATISTIC_KIND_DECHIST, InvalidOid,
                                 ATTSTATSSLOT_NUMBERS))
            {
                if (sslot.nnumbers > 0)
                    nelem = clamp_row_est(sslot.numbers[sslot.nnumbers - 1]);
                free_attstatsslot(&sslot);
            }
        }
        ReleaseVariableStats(vardata);
 
        if (nelem > 0)
            return nelem;
    }
 
    /* Else use a default guess --- this should match scalararraysel */
    return 10;
}

References ARR_DIMS, ARR_NDIM, ArrayGetNItems(), ATTSTATSSLOT_NUMBERS, clamp_row_est(), DatumGetArrayTypeP, examine_variable(), free_attstatsslot(), get_attstatsslot(), HeapTupleIsValid, InvalidOid, IsA, list_length(), AttStatsSlot::nnumbers, AttStatsSlot::numbers, ReleaseVariableStats, root, VariableStatData::statsTuple, and strip_array_coercion().

Referenced by array_unnest_support(), btcostestimate(), cost_qual_eval_walker(), cost_tidscan(), genericcostestimate(), and gincost_scalararrayopexpr().

estimate_hash_bucket_stats()

void estimate_hash_bucket_stats	(	PlannerInfo *	root,
		Node *	hashkey,
		double	nbuckets,
		Selectivity *	mcv_freq,
		Selectivity *	bucketsize_frac
	)

Definition at line 4390 of file selfuncs.c.

{
    VariableStatData vardata;
    double      estfract,
                ndistinct,
                stanullfrac,
                avgfreq;
    bool        isdefault;
    AttStatsSlot sslot;
 
    examine_variable(root, hashkey, 0, &vardata);
 
    /* Initialize *mcv_freq to "unknown" */
    *mcv_freq = 0.0;
 
    /* Look up the frequency of the most common value, if available */
    if (HeapTupleIsValid(vardata.statsTuple))
    {
        if (get_attstatsslot(&sslot, vardata.statsTuple,
                             STATISTIC_KIND_MCV, InvalidOid,
                             ATTSTATSSLOT_NUMBERS))
        {
            /*
             * The first MCV stat is for the most common value.
             */
            if (sslot.nnumbers > 0)
                *mcv_freq = sslot.numbers[0];
            free_attstatsslot(&sslot);
        }
        else if (get_attstatsslot(&sslot, vardata.statsTuple,
                                  STATISTIC_KIND_HISTOGRAM, InvalidOid,
                                  0))
        {
            /*
             * If there are no recorded MCVs, but we do have a histogram, then
             * assume that ANALYZE determined that the column is unique.
             */
            if (vardata.rel && vardata.rel->rows > 0)
                *mcv_freq = 1.0 / vardata.rel->rows;
        }
    }
 
    /* Get number of distinct values */
    ndistinct = get_variable_numdistinct(&vardata, &isdefault);
 
    /*
     * If ndistinct isn't real, punt.  We normally return 0.1, but if the
     * mcv_freq is known to be even higher than that, use it instead.
     */
    if (isdefault)
    {
        *bucketsize_frac = (Selectivity) Max(0.1, *mcv_freq);
        ReleaseVariableStats(vardata);
        return;
    }
 
    /* Get fraction that are null */
    if (HeapTupleIsValid(vardata.statsTuple))
    {
        Form_pg_statistic stats;
 
        stats = (Form_pg_statistic) GETSTRUCT(vardata.statsTuple);
        stanullfrac = stats->stanullfrac;
    }
    else
        stanullfrac = 0.0;
 
    /* Compute avg freq of all distinct data values in raw relation */
    avgfreq = (1.0 - stanullfrac) / ndistinct;
 
    /*
     * Adjust ndistinct to account for restriction clauses.  Observe we are
     * assuming that the data distribution is affected uniformly by the
     * restriction clauses!
     *
     * XXX Possibly better way, but much more expensive: multiply by
     * selectivity of rel's restriction clauses that mention the target Var.
     */
    if (vardata.rel && vardata.rel->tuples > 0)
    {
        ndistinct *= vardata.rel->rows / vardata.rel->tuples;
        ndistinct = clamp_row_est(ndistinct);
    }
 
    /*
     * Initial estimate of bucketsize fraction is 1/nbuckets as long as the
     * number of buckets is less than the expected number of distinct values;
     * otherwise it is 1/ndistinct.
     */
    if (ndistinct > nbuckets)
        estfract = 1.0 / nbuckets;
    else
        estfract = 1.0 / ndistinct;
 
    /*
     * Adjust estimated bucketsize upward to account for skewed distribution.
     */
    if (avgfreq > 0.0 && *mcv_freq > avgfreq)
        estfract *= *mcv_freq / avgfreq;
 
    /*
     * Clamp bucketsize to sane range (the above adjustment could easily
     * produce an out-of-range result).  We set the lower bound a little above
     * zero, since zero isn't a very sane result.
     */
    if (estfract < 1.0e-6)
        estfract = 1.0e-6;
    else if (estfract > 1.0)
        estfract = 1.0;
 
    *bucketsize_frac = (Selectivity) estfract;
 
    ReleaseVariableStats(vardata);
}

References ATTSTATSSLOT_NUMBERS, clamp_row_est(), examine_variable(), free_attstatsslot(), get_attstatsslot(), get_variable_numdistinct(), GETSTRUCT(), HeapTupleIsValid, InvalidOid, Max, AttStatsSlot::nnumbers, AttStatsSlot::numbers, VariableStatData::rel, ReleaseVariableStats, root, RelOptInfo::rows, VariableStatData::statsTuple, and RelOptInfo::tuples.

Referenced by final_cost_hashjoin().

estimate_hashagg_tablesize()

double estimate_hashagg_tablesize	(	PlannerInfo *	root,
		Path *	path,
		const AggClauseCosts *	agg_costs,
		double	dNumGroups
	)

Definition at line 4521 of file selfuncs.c.

{
    Size        hashentrysize;
 
    hashentrysize = hash_agg_entry_size(list_length(root->aggtransinfos),
                                        path->pathtarget->width,
                                        agg_costs->transitionSpace);
 
    /*
     * Note that this disregards the effect of fill-factor and growth policy
     * of the hash table.  That's probably ok, given that the default
     * fill-factor is relatively high.  It'd be hard to meaningfully factor in
     * "double-in-size" growth policies here.
     */
    return hashentrysize * dNumGroups;
}

References hash_agg_entry_size(), list_length(), root, and AggClauseCosts::transitionSpace.

Referenced by consider_groupingsets_paths().

estimate_multivariate_bucketsize()

List * estimate_multivariate_bucketsize	(	PlannerInfo *	root,
		RelOptInfo *	inner,
		List *	hashclauses,
		Selectivity *	innerbucketsize
	)

Definition at line 4123 of file selfuncs.c.

{
    List       *clauses;
    List       *otherclauses;
    double      ndistinct;
 
    if (list_length(hashclauses) <= 1)
    {
        /*
         * Nothing to do for a single clause.  Could we employ univariate
         * extended stat here?
         */
        return hashclauses;
    }
 
    /* "clauses" is the list of hashclauses we've not dealt with yet */
    clauses = list_copy(hashclauses);
    /* "otherclauses" holds clauses we are going to return to caller */
    otherclauses = NIL;
    /* current estimate of ndistinct */
    ndistinct = 1.0;
    while (clauses != NIL)
    {
        ListCell   *lc;
        int         relid = -1;
        List       *varinfos = NIL;
        List       *origin_rinfos = NIL;
        double      mvndistinct;
        List       *origin_varinfos;
        int         group_relid = -1;
        RelOptInfo *group_rel = NULL;
        ListCell   *lc1,
                   *lc2;
 
        /*
         * Find clauses, referencing the same single base relation and try to
         * estimate such a group with extended statistics.  Create varinfo for
         * an approved clause, push it to otherclauses, if it can't be
         * estimated here or ignore to process at the next iteration.
         */
        foreach(lc, clauses)
        {
            RestrictInfo *rinfo = lfirst_node(RestrictInfo, lc);
            Node       *expr;
            Relids      relids;
            GroupVarInfo *varinfo;
 
            /*
             * Find the inner side of the join, which we need to estimate the
             * number of buckets.  Use outer_is_left because the
             * clause_sides_match_join routine has called on hash clauses.
             */
            relids = rinfo->outer_is_left ?
                rinfo->right_relids : rinfo->left_relids;
            expr = rinfo->outer_is_left ?
                get_rightop(rinfo->clause) : get_leftop(rinfo->clause);
 
            if (bms_get_singleton_member(relids, &relid) &&
                root->simple_rel_array[relid]->statlist != NIL)
            {
                bool        is_duplicate = false;
 
                /*
                 * This inner-side expression references only one relation.
                 * Extended statistics on this clause can exist.
                 */
                if (group_relid < 0)
                {
                    RangeTblEntry *rte = root->simple_rte_array[relid];
 
                    if (!rte || (rte->relkind != RELKIND_RELATION &&
                                 rte->relkind != RELKIND_MATVIEW &&
                                 rte->relkind != RELKIND_FOREIGN_TABLE &&
                                 rte->relkind != RELKIND_PARTITIONED_TABLE))
                    {
                        /* Extended statistics can't exist in principle */
                        otherclauses = lappend(otherclauses, rinfo);
                        clauses = foreach_delete_current(clauses, lc);
                        continue;
                    }
 
                    group_relid = relid;
                    group_rel = root->simple_rel_array[relid];
                }
                else if (group_relid != relid)
                {
                    /*
                     * Being in the group forming state we don't need other
                     * clauses.
                     */
                    continue;
                }
 
                /*
                 * We're going to add the new clause to the varinfos list.  We
                 * might re-use add_unique_group_var(), but we don't do so for
                 * two reasons.
                 *
                 * 1) We must keep the origin_rinfos list ordered exactly the
                 * same way as varinfos.
                 *
                 * 2) add_unique_group_var() is designed for
                 * estimate_num_groups(), where a larger number of groups is
                 * worse.   While estimating the number of hash buckets, we
                 * have the opposite: a lesser number of groups is worse.
                 * Therefore, we don't have to remove "known equal" vars: the
                 * removed var may valuably contribute to the multivariate
                 * statistics to grow the number of groups.
                 */
 
                /*
                 * Clear nullingrels to correctly match hash keys.  See
                 * add_unique_group_var()'s comment for details.
                 */
                expr = remove_nulling_relids(expr, root->outer_join_rels, NULL);
 
                /*
                 * Detect and exclude exact duplicates from the list of hash
                 * keys (like add_unique_group_var does).
                 */
                foreach(lc1, varinfos)
                {
                    varinfo = (GroupVarInfo *) lfirst(lc1);
 
                    if (!equal(expr, varinfo->var))
                        continue;
 
                    is_duplicate = true;
                    break;
                }
 
                if (is_duplicate)
                {
                    /*
                     * Skip exact duplicates. Adding them to the otherclauses
                     * list also doesn't make sense.
                     */
                    continue;
                }
 
                /*
                 * Initialize GroupVarInfo.  We only use it to call
                 * estimate_multivariate_ndistinct(), which doesn't care about
                 * ndistinct and isdefault fields.  Thus, skip these fields.
                 */
                varinfo = palloc0_object(GroupVarInfo);
                varinfo->var = expr;
                varinfo->rel = root->simple_rel_array[relid];
                varinfos = lappend(varinfos, varinfo);
 
                /*
                 * Remember the link to RestrictInfo for the case the clause
                 * is failed to be estimated.
                 */
                origin_rinfos = lappend(origin_rinfos, rinfo);
            }
            else
            {
                /* This clause can't be estimated with extended statistics */
                otherclauses = lappend(otherclauses, rinfo);
            }
 
            clauses = foreach_delete_current(clauses, lc);
        }
 
        if (list_length(varinfos) < 2)
        {
            /*
             * Multivariate statistics doesn't apply to single columns except
             * for expressions, but it has not been implemented yet.
             */
            otherclauses = list_concat(otherclauses, origin_rinfos);
            list_free_deep(varinfos);
            list_free(origin_rinfos);
            continue;
        }
 
        Assert(group_rel != NULL);
 
        /* Employ the extended statistics. */
        origin_varinfos = varinfos;
        for (;;)
        {
            bool        estimated = estimate_multivariate_ndistinct(root,
                                                                    group_rel,
                                                                    &varinfos,
                                                                    &mvndistinct);
 
            if (!estimated)
                break;
 
            /*
             * We've got an estimation.  Use ndistinct value in a consistent
             * way - according to the caller's logic (see
             * final_cost_hashjoin).
             */
            if (ndistinct < mvndistinct)
                ndistinct = mvndistinct;
            Assert(ndistinct >= 1.0);
        }
 
        Assert(list_length(origin_varinfos) == list_length(origin_rinfos));
 
        /* Collect unmatched clauses as otherclauses. */
        forboth(lc1, origin_varinfos, lc2, origin_rinfos)
        {
            GroupVarInfo *vinfo = lfirst(lc1);
 
            if (!list_member_ptr(varinfos, vinfo))
                /* Already estimated */
                continue;
 
            /* Can't be estimated here - push to the returning list */
            otherclauses = lappend(otherclauses, lfirst(lc2));
        }
    }
 
    *innerbucketsize = 1.0 / ndistinct;
    return otherclauses;
}

References Assert(), bms_get_singleton_member(), RestrictInfo::clause, equal(), estimate_multivariate_ndistinct(), forboth, foreach_delete_current, get_leftop(), get_rightop(), lappend(), lfirst, lfirst_node, list_concat(), list_copy(), list_free(), list_free_deep(), list_length(), list_member_ptr(), NIL, palloc0_object, GroupVarInfo::rel, remove_nulling_relids(), root, and GroupVarInfo::var.

Referenced by final_cost_hashjoin().

estimate_num_groups()

double estimate_num_groups	(	PlannerInfo *	root,
		List *	groupExprs,
		double	input_rows,
		List **	pgset,
		EstimationInfo *	estinfo
	)

Definition at line 3771 of file selfuncs.c.

{
    List       *varinfos = NIL;
    double      srf_multiplier = 1.0;
    double      numdistinct;
    ListCell   *l;
    int         i;
 
    /* Zero the estinfo output parameter, if non-NULL */
    if (estinfo != NULL)
        memset(estinfo, 0, sizeof(EstimationInfo));
 
    /*
     * We don't ever want to return an estimate of zero groups, as that tends
     * to lead to division-by-zero and other unpleasantness.  The input_rows
     * estimate is usually already at least 1, but clamp it just in case it
     * isn't.
     */
    input_rows = clamp_row_est(input_rows);
 
    /*
     * If no grouping columns, there's exactly one group.  (This can't happen
     * for normal cases with GROUP BY or DISTINCT, but it is possible for
     * corner cases with set operations.)
     */
    if (groupExprs == NIL || (pgset && *pgset == NIL))
        return 1.0;
 
    /*
     * Count groups derived from boolean grouping expressions.  For other
     * expressions, find the unique Vars used, treating an expression as a Var
     * if we can find stats for it.  For each one, record the statistical
     * estimate of number of distinct values (total in its table, without
     * regard for filtering).
     */
    numdistinct = 1.0;
 
    i = 0;
    foreach(l, groupExprs)
    {
        Node       *groupexpr = (Node *) lfirst(l);
        double      this_srf_multiplier;
        VariableStatData vardata;
        List       *varshere;
        ListCell   *l2;
 
        /* is expression in this grouping set? */
        if (pgset && !list_member_int(*pgset, i++))
            continue;
 
        /*
         * Set-returning functions in grouping columns are a bit problematic.
         * The code below will effectively ignore their SRF nature and come up
         * with a numdistinct estimate as though they were scalar functions.
         * We compensate by scaling up the end result by the largest SRF
         * rowcount estimate.  (This will be an overestimate if the SRF
         * produces multiple copies of any output value, but it seems best to
         * assume the SRF's outputs are distinct.  In any case, it's probably
         * pointless to worry too much about this without much better
         * estimates for SRF output rowcounts than we have today.)
         */
        this_srf_multiplier = expression_returns_set_rows(root, groupexpr);
        if (srf_multiplier < this_srf_multiplier)
            srf_multiplier = this_srf_multiplier;
 
        /* Short-circuit for expressions returning boolean */
        if (exprType(groupexpr) == BOOLOID)
        {
            numdistinct *= 2.0;
            continue;
        }
 
        /*
         * If examine_variable is able to deduce anything about the GROUP BY
         * expression, treat it as a single variable even if it's really more
         * complicated.
         *
         * XXX This has the consequence that if there's a statistics object on
         * the expression, we don't split it into individual Vars. This
         * affects our selection of statistics in
         * estimate_multivariate_ndistinct, because it's probably better to
         * use more accurate estimate for each expression and treat them as
         * independent, than to combine estimates for the extracted variables
         * when we don't know how that relates to the expressions.
         */
        examine_variable(root, groupexpr, 0, &vardata);
        if (HeapTupleIsValid(vardata.statsTuple) || vardata.isunique)
        {
            varinfos = add_unique_group_var(root, varinfos,
                                            groupexpr, &vardata);
            ReleaseVariableStats(vardata);
            continue;
        }
        ReleaseVariableStats(vardata);
 
        /*
         * Else pull out the component Vars.  Handle PlaceHolderVars by
         * recursing into their arguments (effectively assuming that the
         * PlaceHolderVar doesn't change the number of groups, which boils
         * down to ignoring the possible addition of nulls to the result set).
         */
        varshere = pull_var_clause(groupexpr,
                                   PVC_RECURSE_AGGREGATES |
                                   PVC_RECURSE_WINDOWFUNCS |
                                   PVC_RECURSE_PLACEHOLDERS);
 
        /*
         * If we find any variable-free GROUP BY item, then either it is a
         * constant (and we can ignore it) or it contains a volatile function;
         * in the latter case we punt and assume that each input row will
         * yield a distinct group.
         */
        if (varshere == NIL)
        {
            if (contain_volatile_functions(groupexpr))
                return input_rows;
            continue;
        }
 
        /*
         * Else add variables to varinfos list
         */
        foreach(l2, varshere)
        {
            Node       *var = (Node *) lfirst(l2);
 
            examine_variable(root, var, 0, &vardata);
            varinfos = add_unique_group_var(root, varinfos, var, &vardata);
            ReleaseVariableStats(vardata);
        }
    }
 
    /*
     * If now no Vars, we must have an all-constant or all-boolean GROUP BY
     * list.
     */
    if (varinfos == NIL)
    {
        /* Apply SRF multiplier as we would do in the long path */
        numdistinct *= srf_multiplier;
        /* Round off */
        numdistinct = ceil(numdistinct);
        /* Guard against out-of-range answers */
        if (numdistinct > input_rows)
            numdistinct = input_rows;
        if (numdistinct < 1.0)
            numdistinct = 1.0;
        return numdistinct;
    }
 
    /*
     * Group Vars by relation and estimate total numdistinct.
     *
     * For each iteration of the outer loop, we process the frontmost Var in
     * varinfos, plus all other Vars in the same relation.  We remove these
     * Vars from the newvarinfos list for the next iteration. This is the
     * easiest way to group Vars of same rel together.
     */
    do
    {
        GroupVarInfo *varinfo1 = (GroupVarInfo *) linitial(varinfos);
        RelOptInfo *rel = varinfo1->rel;
        double      reldistinct = 1;
        double      relmaxndistinct = reldistinct;
        int         relvarcount = 0;
        List       *newvarinfos = NIL;
        List       *relvarinfos = NIL;
 
        /*
         * Split the list of varinfos in two - one for the current rel, one
         * for remaining Vars on other rels.
         */
        relvarinfos = lappend(relvarinfos, varinfo1);
        for_each_from(l, varinfos, 1)
        {
            GroupVarInfo *varinfo2 = (GroupVarInfo *) lfirst(l);
 
            if (varinfo2->rel == varinfo1->rel)
            {
                /* varinfos on current rel */
                relvarinfos = lappend(relvarinfos, varinfo2);
            }
            else
            {
                /* not time to process varinfo2 yet */
                newvarinfos = lappend(newvarinfos, varinfo2);
            }
        }
 
        /*
         * Get the numdistinct estimate for the Vars of this rel.  We
         * iteratively search for multivariate n-distinct with maximum number
         * of vars; assuming that each var group is independent of the others,
         * we multiply them together.  Any remaining relvarinfos after no more
         * multivariate matches are found are assumed independent too, so
         * their individual ndistinct estimates are multiplied also.
         *
         * While iterating, count how many separate numdistinct values we
         * apply.  We apply a fudge factor below, but only if we multiplied
         * more than one such values.
         */
        while (relvarinfos)
        {
            double      mvndistinct;
 
            if (estimate_multivariate_ndistinct(root, rel, &relvarinfos,
                                                &mvndistinct))
            {
                reldistinct *= mvndistinct;
                if (relmaxndistinct < mvndistinct)
                    relmaxndistinct = mvndistinct;
                relvarcount++;
            }
            else
            {
                foreach(l, relvarinfos)
                {
                    GroupVarInfo *varinfo2 = (GroupVarInfo *) lfirst(l);
 
                    reldistinct *= varinfo2->ndistinct;
                    if (relmaxndistinct < varinfo2->ndistinct)
                        relmaxndistinct = varinfo2->ndistinct;
                    relvarcount++;
 
                    /*
                     * When varinfo2's isdefault is set then we'd better set
                     * the SELFLAG_USED_DEFAULT bit in the EstimationInfo.
                     */
                    if (estinfo != NULL && varinfo2->isdefault)
                        estinfo->flags |= SELFLAG_USED_DEFAULT;
                }
 
                /* we're done with this relation */
                relvarinfos = NIL;
            }
        }
 
        /*
         * Sanity check --- don't divide by zero if empty relation.
         */
        Assert(IS_SIMPLE_REL(rel));
        if (rel->tuples > 0)
        {
            /*
             * Clamp to size of rel, or size of rel / 10 if multiple Vars. The
             * fudge factor is because the Vars are probably correlated but we
             * don't know by how much.  We should never clamp to less than the
             * largest ndistinct value for any of the Vars, though, since
             * there will surely be at least that many groups.
             */
            double      clamp = rel->tuples;
 
            if (relvarcount > 1)
            {
                clamp *= 0.1;
                if (clamp < relmaxndistinct)
                {
                    clamp = relmaxndistinct;
                    /* for sanity in case some ndistinct is too large: */
                    if (clamp > rel->tuples)
                        clamp = rel->tuples;
                }
            }
            if (reldistinct > clamp)
                reldistinct = clamp;
 
            /*
             * Update the estimate based on the restriction selectivity,
             * guarding against division by zero when reldistinct is zero.
             * Also skip this if we know that we are returning all rows.
             */
            if (reldistinct > 0 && rel->rows < rel->tuples)
            {
                /*
                 * Given a table containing N rows with n distinct values in a
                 * uniform distribution, if we select p rows at random then
                 * the expected number of distinct values selected is
                 *
                 * n * (1 - product((N-N/n-i)/(N-i), i=0..p-1))
                 *
                 * = n * (1 - (N-N/n)! / (N-N/n-p)! * (N-p)! / N!)
                 *
                 * See "Approximating block accesses in database
                 * organizations", S. B. Yao, Communications of the ACM,
                 * Volume 20 Issue 4, April 1977 Pages 260-261.
                 *
                 * Alternatively, re-arranging the terms from the factorials,
                 * this may be written as
                 *
                 * n * (1 - product((N-p-i)/(N-i), i=0..N/n-1))
                 *
                 * This form of the formula is more efficient to compute in
                 * the common case where p is larger than N/n.  Additionally,
                 * as pointed out by Dell'Era, if i << N for all terms in the
                 * product, it can be approximated by
                 *
                 * n * (1 - ((N-p)/N)^(N/n))
                 *
                 * See "Expected distinct values when selecting from a bag
                 * without replacement", Alberto Dell'Era,
                 * http://www.adellera.it/investigations/distinct_balls/.
                 *
                 * The condition i << N is equivalent to n >> 1, so this is a
                 * good approximation when the number of distinct values in
                 * the table is large.  It turns out that this formula also
                 * works well even when n is small.
                 */
                reldistinct *=
                    (1 - pow((rel->tuples - rel->rows) / rel->tuples,
                             rel->tuples / reldistinct));
            }
            reldistinct = clamp_row_est(reldistinct);
 
            /*
             * Update estimate of total distinct groups.
             */
            numdistinct *= reldistinct;
        }
 
        varinfos = newvarinfos;
    } while (varinfos != NIL);
 
    /* Now we can account for the effects of any SRFs */
    numdistinct *= srf_multiplier;
 
    /* Round off */
    numdistinct = ceil(numdistinct);
 
    /* Guard against out-of-range answers */
    if (numdistinct > input_rows)
        numdistinct = input_rows;
    if (numdistinct < 1.0)
        numdistinct = 1.0;
 
    return numdistinct;
}

References add_unique_group_var(), Assert(), clamp_row_est(), contain_volatile_functions(), estimate_multivariate_ndistinct(), examine_variable(), expression_returns_set_rows(), exprType(), EstimationInfo::flags, for_each_from, HeapTupleIsValid, i, IS_SIMPLE_REL, GroupVarInfo::isdefault, VariableStatData::isunique, lappend(), lfirst, linitial, list_member_int(), GroupVarInfo::ndistinct, NIL, pull_var_clause(), PVC_RECURSE_AGGREGATES, PVC_RECURSE_PLACEHOLDERS, PVC_RECURSE_WINDOWFUNCS, GroupVarInfo::rel, ReleaseVariableStats, root, RelOptInfo::rows, SELFLAG_USED_DEFAULT, VariableStatData::statsTuple, and RelOptInfo::tuples.

Referenced by adjust_rowcount_for_semijoins(), build_setop_child_paths(), cost_incremental_sort(), cost_memoize_rescan(), create_final_distinct_paths(), create_final_unique_paths(), create_partial_distinct_paths(), create_partial_unique_paths(), create_rel_agg_info(), estimate_path_cost_size(), generate_grouped_paths(), generate_union_paths(), get_number_of_groups(), and get_windowclause_startup_tuples().

examine_variable()

void examine_variable	(	PlannerInfo *	root,
		Node *	node,
		int	varRelid,
		VariableStatData *	vardata
	)

Definition at line 5636 of file selfuncs.c.

{
    Node       *basenode;
    Relids      varnos;
    Relids      basevarnos;
    RelOptInfo *onerel;
 
    /* Make sure we don't return dangling pointers in vardata */
    MemSet(vardata, 0, sizeof(VariableStatData));
 
    /* Save the exposed type of the expression */
    vardata->vartype = exprType(node);
 
    /*
     * PlaceHolderVars are transparent for the purpose of statistics lookup;
     * they do not alter the value distribution of the underlying expression.
     * However, they can obscure the structure, preventing us from recognizing
     * matches to base columns, index expressions, or extended statistics.  So
     * strip them out first.
     */
    basenode = strip_all_phvs_deep(root, node);
 
    /*
     * Look inside any binary-compatible relabeling.  We need to handle nested
     * RelabelType nodes here, because the prior stripping of PlaceHolderVars
     * may have brought separate RelabelTypes into adjacency.
     */
    while (IsA(basenode, RelabelType))
        basenode = (Node *) ((RelabelType *) basenode)->arg;
 
    /* Fast path for a simple Var */
    if (IsA(basenode, Var) &&
        (varRelid == 0 || varRelid == ((Var *) basenode)->varno))
    {
        Var        *var = (Var *) basenode;
 
        /* Set up result fields other than the stats tuple */
        vardata->var = basenode;    /* return Var without phvs or relabeling */
        vardata->rel = find_base_rel(root, var->varno);
        vardata->atttype = var->vartype;
        vardata->atttypmod = var->vartypmod;
        vardata->isunique = has_unique_index(vardata->rel, var->varattno);
 
        /* Try to locate some stats */
        examine_simple_variable(root, var, vardata);
 
        return;
    }
 
    /*
     * Okay, it's a more complicated expression.  Determine variable
     * membership.  Note that when varRelid isn't zero, only vars of that
     * relation are considered "real" vars.
     */
    varnos = pull_varnos(root, basenode);
    basevarnos = bms_difference(varnos, root->outer_join_rels);
 
    onerel = NULL;
 
    if (bms_is_empty(basevarnos))
    {
        /* No Vars at all ... must be pseudo-constant clause */
    }
    else
    {
        int         relid;
 
        /* Check if the expression is in vars of a single base relation */
        if (bms_get_singleton_member(basevarnos, &relid))
        {
            if (varRelid == 0 || varRelid == relid)
            {
                onerel = find_base_rel(root, relid);
                vardata->rel = onerel;
                node = basenode;    /* strip any phvs or relabeling */
            }
            /* else treat it as a constant */
        }
        else
        {
            /* varnos has multiple relids */
            if (varRelid == 0)
            {
                /* treat it as a variable of a join relation */
                vardata->rel = find_join_rel(root, varnos);
                node = basenode;    /* strip any phvs or relabeling */
            }
            else if (bms_is_member(varRelid, varnos))
            {
                /* ignore the vars belonging to other relations */
                vardata->rel = find_base_rel(root, varRelid);
                node = basenode;    /* strip any phvs or relabeling */
                /* note: no point in expressional-index search here */
            }
            /* else treat it as a constant */
        }
    }
 
    bms_free(basevarnos);
 
    vardata->var = node;
    vardata->atttype = exprType(node);
    vardata->atttypmod = exprTypmod(node);
 
    if (onerel)
    {
        /*
         * We have an expression in vars of a single relation.  Try to match
         * it to expressional index columns, in hopes of finding some
         * statistics.
         *
         * Note that we consider all index columns including INCLUDE columns,
         * since there could be stats for such columns.  But the test for
         * uniqueness needs to be warier.
         *
         * XXX it's conceivable that there are multiple matches with different
         * index opfamilies; if so, we need to pick one that matches the
         * operator we are estimating for.  FIXME later.
         */
        ListCell   *ilist;
        ListCell   *slist;
 
        /*
         * The nullingrels bits within the expression could prevent us from
         * matching it to expressional index columns or to the expressions in
         * extended statistics.  So strip them out first.
         */
        if (bms_overlap(varnos, root->outer_join_rels))
            node = remove_nulling_relids(node, root->outer_join_rels, NULL);
 
        foreach(ilist, onerel->indexlist)
        {
            IndexOptInfo *index = (IndexOptInfo *) lfirst(ilist);
            ListCell   *indexpr_item;
            int         pos;
 
            indexpr_item = list_head(index->indexprs);
            if (indexpr_item == NULL)
                continue;       /* no expressions here... */
 
            for (pos = 0; pos < index->ncolumns; pos++)
            {
                if (index->indexkeys[pos] == 0)
                {
                    Node       *indexkey;
 
                    if (indexpr_item == NULL)
                        elog(ERROR, "too few entries in indexprs list");
                    indexkey = (Node *) lfirst(indexpr_item);
                    if (indexkey && IsA(indexkey, RelabelType))
                        indexkey = (Node *) ((RelabelType *) indexkey)->arg;
                    if (equal(node, indexkey))
                    {
                        /*
                         * Found a match ... is it a unique index? Tests here
                         * should match has_unique_index().
                         */
                        if (index->unique &&
                            index->nkeycolumns == 1 &&
                            pos == 0 &&
                            (index->indpred == NIL || index->predOK))
                            vardata->isunique = true;
 
                        /*
                         * Has it got stats?  We only consider stats for
                         * non-partial indexes, since partial indexes probably
                         * don't reflect whole-relation statistics; the above
                         * check for uniqueness is the only info we take from
                         * a partial index.
                         *
                         * An index stats hook, however, must make its own
                         * decisions about what to do with partial indexes.
                         */
                        if (get_index_stats_hook &&
                            (*get_index_stats_hook) (root, index->indexoid,
                                                     pos + 1, vardata))
                        {
                            /*
                             * The hook took control of acquiring a stats
                             * tuple.  If it did supply a tuple, it'd better
                             * have supplied a freefunc.
                             */
                            if (HeapTupleIsValid(vardata->statsTuple) &&
                                !vardata->freefunc)
                                elog(ERROR, "no function provided to release variable stats with");
                        }
                        else if (index->indpred == NIL)
                        {
                            vardata->statsTuple =
                                SearchSysCache3(STATRELATTINH,
                                                ObjectIdGetDatum(index->indexoid),
                                                Int16GetDatum(pos + 1),
                                                BoolGetDatum(false));
                            vardata->freefunc = ReleaseSysCache;
 
                            if (HeapTupleIsValid(vardata->statsTuple))
                            {
                                /*
                                 * Test if user has permission to access all
                                 * rows from the index's table.
                                 *
                                 * For simplicity, we insist on the whole
                                 * table being selectable, rather than trying
                                 * to identify which column(s) the index
                                 * depends on.
                                 *
                                 * Note that for an inheritance child,
                                 * permissions are checked on the inheritance
                                 * root parent, and whole-table select
                                 * privilege on the parent doesn't quite
                                 * guarantee that the user could read all
                                 * columns of the child.  But in practice it's
                                 * unlikely that any interesting security
                                 * violation could result from allowing access
                                 * to the expression index's stats, so we
                                 * allow it anyway.  See similar code in
                                 * examine_simple_variable() for additional
                                 * comments.
                                 */
                                vardata->acl_ok =
                                    all_rows_selectable(root,
                                                        index->rel->relid,
                                                        NULL);
                            }
                            else
                            {
                                /* suppress leakproofness checks later */
                                vardata->acl_ok = true;
                            }
                        }
                        if (vardata->statsTuple)
                            break;
                    }
                    indexpr_item = lnext(index->indexprs, indexpr_item);
                }
            }
            if (vardata->statsTuple)
                break;
        }
 
        /*
         * Search extended statistics for one with a matching expression.
         * There might be multiple ones, so just grab the first one. In the
         * future, we might consider the statistics target (and pick the most
         * accurate statistics) and maybe some other parameters.
         */
        foreach(slist, onerel->statlist)
        {
            StatisticExtInfo *info = (StatisticExtInfo *) lfirst(slist);
            RangeTblEntry *rte = planner_rt_fetch(onerel->relid, root);
            ListCell   *expr_item;
            int         pos;
 
            /*
             * Stop once we've found statistics for the expression (either
             * from extended stats, or for an index in the preceding loop).
             */
            if (vardata->statsTuple)
                break;
 
            /* skip stats without per-expression stats */
            if (info->kind != STATS_EXT_EXPRESSIONS)
                continue;
 
            /* skip stats with mismatching stxdinherit value */
            if (info->inherit != rte->inh)
                continue;
 
            pos = 0;
            foreach(expr_item, info->exprs)
            {
                Node       *expr = (Node *) lfirst(expr_item);
 
                Assert(expr);
 
                /* strip RelabelType before comparing it */
                if (expr && IsA(expr, RelabelType))
                    expr = (Node *) ((RelabelType *) expr)->arg;
 
                /* found a match, see if we can extract pg_statistic row */
                if (equal(node, expr))
                {
                    /*
                     * XXX Not sure if we should cache the tuple somewhere.
                     * Now we just create a new copy every time.
                     */
                    vardata->statsTuple =
                        statext_expressions_load(info->statOid, rte->inh, pos);
 
                    vardata->freefunc = ReleaseDummy;
 
                    /*
                     * Test if user has permission to access all rows from the
                     * table.
                     *
                     * For simplicity, we insist on the whole table being
                     * selectable, rather than trying to identify which
                     * column(s) the statistics object depends on.
                     *
                     * Note that for an inheritance child, permissions are
                     * checked on the inheritance root parent, and whole-table
                     * select privilege on the parent doesn't quite guarantee
                     * that the user could read all columns of the child.  But
                     * in practice it's unlikely that any interesting security
                     * violation could result from allowing access to the
                     * expression stats, so we allow it anyway.  See similar
                     * code in examine_simple_variable() for additional
                     * comments.
                     */
                    vardata->acl_ok = all_rows_selectable(root,
                                                          onerel->relid,
                                                          NULL);
 
                    break;
                }
 
                pos++;
            }
        }
    }
 
    bms_free(varnos);
}

References VariableStatData::acl_ok, all_rows_selectable(), arg, Assert(), VariableStatData::atttype, VariableStatData::atttypmod, bms_difference(), bms_free(), bms_get_singleton_member(), bms_is_empty, bms_is_member(), bms_overlap(), BoolGetDatum(), elog, equal(), ERROR, examine_simple_variable(), StatisticExtInfo::exprs, exprType(), exprTypmod(), find_base_rel(), find_join_rel(), VariableStatData::freefunc, get_index_stats_hook, has_unique_index(), HeapTupleIsValid, if(), RelOptInfo::indexlist, RangeTblEntry::inh, StatisticExtInfo::inherit, Int16GetDatum(), IsA, VariableStatData::isunique, StatisticExtInfo::kind, lfirst, list_head(), lnext(), MemSet, NIL, ObjectIdGetDatum(), planner_rt_fetch, pull_varnos(), VariableStatData::rel, ReleaseDummy(), ReleaseSysCache(), RelOptInfo::relid, remove_nulling_relids(), root, SearchSysCache3(), statext_expressions_load(), RelOptInfo::statlist, StatisticExtInfo::statOid, VariableStatData::statsTuple, strip_all_phvs_deep(), VariableStatData::var, Var::varattno, Var::varno, and VariableStatData::vartype.

Referenced by booltestsel(), boolvarsel(), estimate_array_length(), estimate_hash_bucket_stats(), estimate_num_groups(), get_join_variables(), get_restriction_variable(), mergejoinscansel(), nulltestsel(), and scalararraysel_containment().

generic_restriction_selectivity()

double generic_restriction_selectivity	(	PlannerInfo *	root,
		Oid	oproid,
		Oid	collation,
		List *	args,
		int	varRelid,
		double	default_selectivity
	)

Definition at line 987 of file selfuncs.c.

{
    double      selec;
    VariableStatData vardata;
    Node       *other;
    bool        varonleft;
 
    /*
     * If expression is not variable OP something or something OP variable,
     * then punt and return the default estimate.
     */
    if (!get_restriction_variable(root, args, varRelid,
                                  &vardata, &other, &varonleft))
        return default_selectivity;
 
    /*
     * If the something is a NULL constant, assume operator is strict and
     * return zero, ie, operator will never return TRUE.
     */
    if (IsA(other, Const) &&
        ((Const *) other)->constisnull)
    {
        ReleaseVariableStats(vardata);
        return 0.0;
    }
 
    if (IsA(other, Const))
    {
        /* Variable is being compared to a known non-null constant */
        Datum       constval = ((Const *) other)->constvalue;
        FmgrInfo    opproc;
        double      mcvsum;
        double      mcvsel;
        double      nullfrac;
        int         hist_size;
 
        fmgr_info(get_opcode(oproid), &opproc);
 
        /*
         * Calculate the selectivity for the column's most common values.
         */
        mcvsel = mcv_selectivity(&vardata, &opproc, collation,
                                 constval, varonleft,
                                 &mcvsum);
 
        /*
         * If the histogram is large enough, see what fraction of it matches
         * the query, and assume that's representative of the non-MCV
         * population.  Otherwise use the default selectivity for the non-MCV
         * population.
         */
        selec = histogram_selectivity(&vardata, &opproc, collation,
                                      constval, varonleft,
                                      10, 1, &hist_size);
        if (selec < 0)
        {
            /* Nope, fall back on default */
            selec = default_selectivity;
        }
        else if (hist_size < 100)
        {
            /*
             * For histogram sizes from 10 to 100, we combine the histogram
             * and default selectivities, putting increasingly more trust in
             * the histogram for larger sizes.
             */
            double      hist_weight = hist_size / 100.0;
 
            selec = selec * hist_weight +
                default_selectivity * (1.0 - hist_weight);
        }
 
        /* In any case, don't believe extremely small or large estimates. */
        if (selec < 0.0001)
            selec = 0.0001;
        else if (selec > 0.9999)
            selec = 0.9999;
 
        /* Don't forget to account for nulls. */
        if (HeapTupleIsValid(vardata.statsTuple))
            nullfrac = ((Form_pg_statistic) GETSTRUCT(vardata.statsTuple))->stanullfrac;
        else
            nullfrac = 0.0;
 
        /*
         * Now merge the results from the MCV and histogram calculations,
         * realizing that the histogram covers only the non-null values that
         * are not listed in MCV.
         */
        selec *= 1.0 - nullfrac - mcvsum;
        selec += mcvsel;
    }
    else
    {
        /* Comparison value is not constant, so we can't do anything */
        selec = default_selectivity;
    }
 
    ReleaseVariableStats(vardata);
 
    /* result should be in range, but make sure... */
    CLAMP_PROBABILITY(selec);
 
    return selec;
}

References generate_unaccent_rules::args, CLAMP_PROBABILITY, fmgr_info(), get_opcode(), get_restriction_variable(), GETSTRUCT(), HeapTupleIsValid, histogram_selectivity(), IsA, mcv_selectivity(), ReleaseVariableStats, root, and VariableStatData::statsTuple.

Referenced by ltreeparentsel(), and matchingsel().

genericcostestimate()

void genericcostestimate	(	PlannerInfo *	root,
		IndexPath *	path,
		double	loop_count,
		GenericCosts *	costs
	)

Definition at line 7395 of file selfuncs.c.

{
    IndexOptInfo *index = path->indexinfo;
    List       *indexQuals = get_quals_from_indexclauses(path->indexclauses);
    List       *indexOrderBys = path->indexorderbys;
    Cost        indexStartupCost;
    Cost        indexTotalCost;
    Selectivity indexSelectivity;
    double      indexCorrelation;
    double      numIndexPages;
    double      numIndexTuples;
    double      spc_random_page_cost;
    double      num_sa_scans;
    double      num_outer_scans;
    double      num_scans;
    double      qual_op_cost;
    double      qual_arg_cost;
    List       *selectivityQuals;
    ListCell   *l;
 
    /*
     * If the index is partial, AND the index predicate with the explicitly
     * given indexquals to produce a more accurate idea of the index
     * selectivity.
     */
    selectivityQuals = add_predicate_to_index_quals(index, indexQuals);
 
    /*
     * If caller didn't give us an estimate for ScalarArrayOpExpr index scans,
     * just assume that the number of index descents is the number of distinct
     * combinations of array elements from all of the scan's SAOP clauses.
     */
    num_sa_scans = costs->num_sa_scans;
    if (num_sa_scans < 1)
    {
        num_sa_scans = 1;
        foreach(l, indexQuals)
        {
            RestrictInfo *rinfo = (RestrictInfo *) lfirst(l);
 
            if (IsA(rinfo->clause, ScalarArrayOpExpr))
            {
                ScalarArrayOpExpr *saop = (ScalarArrayOpExpr *) rinfo->clause;
                double      alength = estimate_array_length(root, lsecond(saop->args));
 
                if (alength > 1)
                    num_sa_scans *= alength;
            }
        }
    }
 
    /* Estimate the fraction of main-table tuples that will be visited */
    indexSelectivity = clauselist_selectivity(root, selectivityQuals,
                                              index->rel->relid,
                                              JOIN_INNER,
                                              NULL);
 
    /*
     * If caller didn't give us an estimate, estimate the number of index
     * tuples that will be visited.  We do it in this rather peculiar-looking
     * way in order to get the right answer for partial indexes.
     */
    numIndexTuples = costs->numIndexTuples;
    if (numIndexTuples <= 0.0)
    {
        numIndexTuples = indexSelectivity * index->rel->tuples;
 
        /*
         * The above calculation counts all the tuples visited across all
         * scans induced by ScalarArrayOpExpr nodes.  We want to consider the
         * average per-indexscan number, so adjust.  This is a handy place to
         * round to integer, too.  (If caller supplied tuple estimate, it's
         * responsible for handling these considerations.)
         */
        numIndexTuples = rint(numIndexTuples / num_sa_scans);
    }
 
    /*
     * We can bound the number of tuples by the index size in any case. Also,
     * always estimate at least one tuple is touched, even when
     * indexSelectivity estimate is tiny.
     */
    if (numIndexTuples > index->tuples)
        numIndexTuples = index->tuples;
    if (numIndexTuples < 1.0)
        numIndexTuples = 1.0;
 
    /*
     * Estimate the number of index pages that will be retrieved.
     *
     * We use the simplistic method of taking a pro-rata fraction of the total
     * number of index pages.  In effect, this counts only leaf pages and not
     * any overhead such as index metapage or upper tree levels.
     *
     * In practice access to upper index levels is often nearly free because
     * those tend to stay in cache under load; moreover, the cost involved is
     * highly dependent on index type.  We therefore ignore such costs here
     * and leave it to the caller to add a suitable charge if needed.
     */
    if (index->pages > 1 && index->tuples > 1)
        numIndexPages = ceil(numIndexTuples * index->pages / index->tuples);
    else
        numIndexPages = 1.0;
 
    /* fetch estimated page cost for tablespace containing index */
    get_tablespace_page_costs(index->reltablespace,
                              &spc_random_page_cost,
                              NULL);
 
    /*
     * Now compute the disk access costs.
     *
     * The above calculations are all per-index-scan.  However, if we are in a
     * nestloop inner scan, we can expect the scan to be repeated (with
     * different search keys) for each row of the outer relation.  Likewise,
     * ScalarArrayOpExpr quals result in multiple index scans.  This creates
     * the potential for cache effects to reduce the number of disk page
     * fetches needed.  We want to estimate the average per-scan I/O cost in
     * the presence of caching.
     *
     * We use the Mackert-Lohman formula (see costsize.c for details) to
     * estimate the total number of page fetches that occur.  While this
     * wasn't what it was designed for, it seems a reasonable model anyway.
     * Note that we are counting pages not tuples anymore, so we take N = T =
     * index size, as if there were one "tuple" per page.
     */
    num_outer_scans = loop_count;
    num_scans = num_sa_scans * num_outer_scans;
 
    if (num_scans > 1)
    {
        double      pages_fetched;
 
        /* total page fetches ignoring cache effects */
        pages_fetched = numIndexPages * num_scans;
 
        /* use Mackert and Lohman formula to adjust for cache effects */
        pages_fetched = index_pages_fetched(pages_fetched,
                                            index->pages,
                                            (double) index->pages,
                                            root);
 
        /*
         * Now compute the total disk access cost, and then report a pro-rated
         * share for each outer scan.  (Don't pro-rate for ScalarArrayOpExpr,
         * since that's internal to the indexscan.)
         */
        indexTotalCost = (pages_fetched * spc_random_page_cost)
            / num_outer_scans;
    }
    else
    {
        /*
         * For a single index scan, we just charge spc_random_page_cost per
         * page touched.
         */
        indexTotalCost = numIndexPages * spc_random_page_cost;
    }
 
    /*
     * CPU cost: any complex expressions in the indexquals will need to be
     * evaluated once at the start of the scan to reduce them to runtime keys
     * to pass to the index AM (see nodeIndexscan.c).  We model the per-tuple
     * CPU costs as cpu_index_tuple_cost plus one cpu_operator_cost per
     * indexqual operator.  Because we have numIndexTuples as a per-scan
     * number, we have to multiply by num_sa_scans to get the correct result
     * for ScalarArrayOpExpr cases.  Similarly add in costs for any index
     * ORDER BY expressions.
     *
     * Note: this neglects the possible costs of rechecking lossy operators.
     * Detecting that that might be needed seems more expensive than it's
     * worth, though, considering all the other inaccuracies here ...
     */
    qual_arg_cost = index_other_operands_eval_cost(root, indexQuals) +
        index_other_operands_eval_cost(root, indexOrderBys);
    qual_op_cost = cpu_operator_cost *
        (list_length(indexQuals) + list_length(indexOrderBys));
 
    indexStartupCost = qual_arg_cost;
    indexTotalCost += qual_arg_cost;
    indexTotalCost += numIndexTuples * num_sa_scans * (cpu_index_tuple_cost + qual_op_cost);
 
    /*
     * Generic assumption about index correlation: there isn't any.
     */
    indexCorrelation = 0.0;
 
    /*
     * Return everything to caller.
     */
    costs->indexStartupCost = indexStartupCost;
    costs->indexTotalCost = indexTotalCost;
    costs->indexSelectivity = indexSelectivity;
    costs->indexCorrelation = indexCorrelation;
    costs->numIndexPages = numIndexPages;
    costs->numIndexTuples = numIndexTuples;
    costs->spc_random_page_cost = spc_random_page_cost;
    costs->num_sa_scans = num_sa_scans;
}

References add_predicate_to_index_quals(), ScalarArrayOpExpr::args, RestrictInfo::clause, clauselist_selectivity(), cpu_index_tuple_cost, cpu_operator_cost, estimate_array_length(), get_quals_from_indexclauses(), get_tablespace_page_costs(), index_other_operands_eval_cost(), index_pages_fetched(), IndexPath::indexclauses, GenericCosts::indexCorrelation, IndexPath::indexinfo, IndexPath::indexorderbys, GenericCosts::indexSelectivity, GenericCosts::indexStartupCost, GenericCosts::indexTotalCost, IsA, JOIN_INNER, lfirst, list_length(), lsecond, GenericCosts::num_sa_scans, GenericCosts::numIndexPages, GenericCosts::numIndexTuples, root, and GenericCosts::spc_random_page_cost.

Referenced by blcostestimate(), btcostestimate(), gistcostestimate(), hashcostestimate(), and spgcostestimate().

get_join_variables()

void get_join_variables	(	PlannerInfo *	root,
		List *	args,
		SpecialJoinInfo *	sjinfo,
		VariableStatData *	vardata1,
		VariableStatData *	vardata2,
		bool *	join_is_reversed
	)

Definition at line 5567 of file selfuncs.c.

{
    Node       *left,
               *right;
 
    if (list_length(args) != 2)
        elog(ERROR, "join operator should take two arguments");
 
    left = (Node *) linitial(args);
    right = (Node *) lsecond(args);
 
    examine_variable(root, left, 0, vardata1);
    examine_variable(root, right, 0, vardata2);
 
    if (vardata1->rel &&
        bms_is_subset(vardata1->rel->relids, sjinfo->syn_righthand))
        *join_is_reversed = true;   /* var1 is on RHS */
    else if (vardata2->rel &&
             bms_is_subset(vardata2->rel->relids, sjinfo->syn_lefthand))
        *join_is_reversed = true;   /* var2 is on LHS */
    else
        *join_is_reversed = false;
}

References generate_unaccent_rules::args, bms_is_subset(), elog, ERROR, examine_variable(), linitial, list_length(), lsecond, VariableStatData::rel, RelOptInfo::relids, root, SpecialJoinInfo::syn_lefthand, and SpecialJoinInfo::syn_righthand.

Referenced by eqjoinsel(), neqjoinsel(), and networkjoinsel().

get_quals_from_indexclauses()

List * get_quals_from_indexclauses

(

List *

indexclauses

)

Definition at line 7311 of file selfuncs.c.

{
    List       *result = NIL;
    ListCell   *lc;
 
    foreach(lc, indexclauses)
    {
        IndexClause *iclause = lfirst_node(IndexClause, lc);
        ListCell   *lc2;
 
        foreach(lc2, iclause->indexquals)
        {
            RestrictInfo *rinfo = lfirst_node(RestrictInfo, lc2);
 
            result = lappend(result, rinfo);
        }
    }
    return result;
}

References IndexClause::indexquals, lappend(), lfirst_node, and NIL.

Referenced by brincostestimate(), genericcostestimate(), and gincostestimate().

get_restriction_variable()

bool get_restriction_variable	(	PlannerInfo *	root,
		List *	args,
		int	varRelid,
		VariableStatData *	vardata,
		Node **	other,
		bool *	varonleft
	)

Definition at line 5507 of file selfuncs.c.

{
    Node       *left,
               *right;
    VariableStatData rdata;
 
    /* Fail if not a binary opclause (probably shouldn't happen) */
    if (list_length(args) != 2)
        return false;
 
    left = (Node *) linitial(args);
    right = (Node *) lsecond(args);
 
    /*
     * Examine both sides.  Note that when varRelid is nonzero, Vars of other
     * relations will be treated as pseudoconstants.
     */
    examine_variable(root, left, varRelid, vardata);
    examine_variable(root, right, varRelid, &rdata);
 
    /*
     * If one side is a variable and the other not, we win.
     */
    if (vardata->rel && rdata.rel == NULL)
    {
        *varonleft = true;
        *other = estimate_expression_value(root, rdata.var);
        /* Assume we need no ReleaseVariableStats(rdata) here */
        return true;
    }
 
    if (vardata->rel == NULL && rdata.rel)
    {
        *varonleft = false;
        *other = estimate_expression_value(root, vardata->var);
        /* Assume we need no ReleaseVariableStats(*vardata) here */
        *vardata = rdata;
        return true;
    }
 
    /* Oops, clause has wrong structure (probably var op var) */
    ReleaseVariableStats(*vardata);
    ReleaseVariableStats(rdata);
 
    return false;
}

References generate_unaccent_rules::args, estimate_expression_value(), examine_variable(), linitial, list_length(), lsecond, VariableStatData::rel, ReleaseVariableStats, root, and VariableStatData::var.

Referenced by _int_matchsel(), arraycontsel(), eqsel_internal(), generic_restriction_selectivity(), multirangesel(), networksel(), patternsel_common(), rangesel(), scalarineqsel_wrapper(), and tsmatchsel().

get_variable_numdistinct()

double get_variable_numdistinct	(	VariableStatData *	vardata,
		bool *	isdefault
	)

Definition at line 6602 of file selfuncs.c.

{
    double      stadistinct;
    double      stanullfrac = 0.0;
    double      ntuples;
 
    *isdefault = false;
 
    /*
     * Determine the stadistinct value to use.  There are cases where we can
     * get an estimate even without a pg_statistic entry, or can get a better
     * value than is in pg_statistic.  Grab stanullfrac too if we can find it
     * (otherwise, assume no nulls, for lack of any better idea).
     */
    if (HeapTupleIsValid(vardata->statsTuple))
    {
        /* Use the pg_statistic entry */
        Form_pg_statistic stats;
 
        stats = (Form_pg_statistic) GETSTRUCT(vardata->statsTuple);
        stadistinct = stats->stadistinct;
        stanullfrac = stats->stanullfrac;
    }
    else if (vardata->vartype == BOOLOID)
    {
        /*
         * Special-case boolean columns: presumably, two distinct values.
         *
         * Are there any other datatypes we should wire in special estimates
         * for?
         */
        stadistinct = 2.0;
    }
    else if (vardata->rel && vardata->rel->rtekind == RTE_VALUES)
    {
        /*
         * If the Var represents a column of a VALUES RTE, assume it's unique.
         * This could of course be very wrong, but it should tend to be true
         * in well-written queries.  We could consider examining the VALUES'
         * contents to get some real statistics; but that only works if the
         * entries are all constants, and it would be pretty expensive anyway.
         */
        stadistinct = -1.0;     /* unique (and all non null) */
    }
    else
    {
        /*
         * We don't keep statistics for system columns, but in some cases we
         * can infer distinctness anyway.
         */
        if (vardata->var && IsA(vardata->var, Var))
        {
            switch (((Var *) vardata->var)->varattno)
            {
                case SelfItemPointerAttributeNumber:
                    stadistinct = -1.0; /* unique (and all non null) */
                    break;
                case TableOidAttributeNumber:
                    stadistinct = 1.0;  /* only 1 value */
                    break;
                default:
                    stadistinct = 0.0;  /* means "unknown" */
                    break;
            }
        }
        else
            stadistinct = 0.0;  /* means "unknown" */
 
        /*
         * XXX consider using estimate_num_groups on expressions?
         */
    }
 
    /*
     * If there is a unique index, DISTINCT or GROUP-BY clause for the
     * variable, assume it is unique no matter what pg_statistic says; the
     * statistics could be out of date, or we might have found a partial
     * unique index that proves the var is unique for this query.  However,
     * we'd better still believe the null-fraction statistic.
     */
    if (vardata->isunique)
        stadistinct = -1.0 * (1.0 - stanullfrac);
 
    /*
     * If we had an absolute estimate, use that.
     */
    if (stadistinct > 0.0)
        return clamp_row_est(stadistinct);
 
    /*
     * Otherwise we need to get the relation size; punt if not available.
     */
    if (vardata->rel == NULL)
    {
        *isdefault = true;
        return DEFAULT_NUM_DISTINCT;
    }
    ntuples = vardata->rel->tuples;
    if (ntuples <= 0.0)
    {
        *isdefault = true;
        return DEFAULT_NUM_DISTINCT;
    }
 
    /*
     * If we had a relative estimate, use that.
     */
    if (stadistinct < 0.0)
        return clamp_row_est(-stadistinct * ntuples);
 
    /*
     * With no data, estimate ndistinct = ntuples if the table is small, else
     * use default.  We use DEFAULT_NUM_DISTINCT as the cutoff for "small" so
     * that the behavior isn't discontinuous.
     */
    if (ntuples < DEFAULT_NUM_DISTINCT)
        return clamp_row_est(ntuples);
 
    *isdefault = true;
    return DEFAULT_NUM_DISTINCT;
}

References clamp_row_est(), DEFAULT_NUM_DISTINCT, GETSTRUCT(), HeapTupleIsValid, IsA, VariableStatData::isunique, VariableStatData::rel, RTE_VALUES, RelOptInfo::rtekind, SelfItemPointerAttributeNumber, VariableStatData::statsTuple, TableOidAttributeNumber, RelOptInfo::tuples, VariableStatData::var, and VariableStatData::vartype.

Referenced by add_unique_group_var(), btcostestimate(), eqjoinsel(), estimate_hash_bucket_stats(), ineq_histogram_selectivity(), var_eq_const(), and var_eq_non_const().

histogram_selectivity()

double histogram_selectivity	(	VariableStatData *	vardata,
		FmgrInfo *	opproc,
		Oid	collation,
		Datum	constval,
		bool	varonleft,
		int	min_hist_size,
		int	n_skip,
		int *	hist_size
	)

Definition at line 896 of file selfuncs.c.

{
    double      result;
    AttStatsSlot sslot;
 
    /* check sanity of parameters */
    Assert(n_skip >= 0);
    Assert(min_hist_size > 2 * n_skip);
 
    if (HeapTupleIsValid(vardata->statsTuple) &&
        statistic_proc_security_check(vardata, opproc->fn_oid) &&
        get_attstatsslot(&sslot, vardata->statsTuple,
                         STATISTIC_KIND_HISTOGRAM, InvalidOid,
                         ATTSTATSSLOT_VALUES))
    {
        *hist_size = sslot.nvalues;
        if (sslot.nvalues >= min_hist_size)
        {
            LOCAL_FCINFO(fcinfo, 2);
            int         nmatch = 0;
            int         i;
 
            /*
             * We invoke the opproc "by hand" so that we won't fail on NULL
             * results.  Such cases won't arise for normal comparison
             * functions, but generic_restriction_selectivity could perhaps be
             * used with operators that can return NULL.  A small side benefit
             * is to not need to re-initialize the fcinfo struct from scratch
             * each time.
             */
            InitFunctionCallInfoData(*fcinfo, opproc, 2, collation,
                                     NULL, NULL);
            fcinfo->args[0].isnull = false;
            fcinfo->args[1].isnull = false;
            /* be careful to apply operator right way 'round */
            if (varonleft)
                fcinfo->args[1].value = constval;
            else
                fcinfo->args[0].value = constval;
 
            for (i = n_skip; i < sslot.nvalues - n_skip; i++)
            {
                Datum       fresult;
 
                if (varonleft)
                    fcinfo->args[0].value = sslot.values[i];
                else
                    fcinfo->args[1].value = sslot.values[i];
                fcinfo->isnull = false;
                fresult = FunctionCallInvoke(fcinfo);
                if (!fcinfo->isnull && DatumGetBool(fresult))
                    nmatch++;
            }
            result = ((double) nmatch) / ((double) (sslot.nvalues - 2 * n_skip));
        }
        else
            result = -1;
        free_attstatsslot(&sslot);
    }
    else
    {
        *hist_size = 0;
        result = -1;
    }
 
    return result;
}

References Assert(), ATTSTATSSLOT_VALUES, DatumGetBool(), FmgrInfo::fn_oid, free_attstatsslot(), FunctionCallInvoke, get_attstatsslot(), HeapTupleIsValid, i, InitFunctionCallInfoData, InvalidOid, LOCAL_FCINFO, AttStatsSlot::nvalues, statistic_proc_security_check(), VariableStatData::statsTuple, and AttStatsSlot::values.

Referenced by generic_restriction_selectivity(), and patternsel_common().

index_other_operands_eval_cost()

Cost index_other_operands_eval_cost	(	PlannerInfo *	root,
		List *	indexquals
	)

Definition at line 7341 of file selfuncs.c.

{
    Cost        qual_arg_cost = 0;
    ListCell   *lc;
 
    foreach(lc, indexquals)
    {
        Expr       *clause = (Expr *) lfirst(lc);
        Node       *other_operand;
        QualCost    index_qual_cost;
 
        /*
         * Index quals will have RestrictInfos, indexorderbys won't.  Look
         * through RestrictInfo if present.
         */
        if (IsA(clause, RestrictInfo))
            clause = ((RestrictInfo *) clause)->clause;
 
        if (IsA(clause, OpExpr))
        {
            OpExpr     *op = (OpExpr *) clause;
 
            other_operand = (Node *) lsecond(op->args);
        }
        else if (IsA(clause, RowCompareExpr))
        {
            RowCompareExpr *rc = (RowCompareExpr *) clause;
 
            other_operand = (Node *) rc->rargs;
        }
        else if (IsA(clause, ScalarArrayOpExpr))
        {
            ScalarArrayOpExpr *saop = (ScalarArrayOpExpr *) clause;
 
            other_operand = (Node *) lsecond(saop->args);
        }
        else if (IsA(clause, NullTest))
        {
            other_operand = NULL;
        }
        else
        {
            elog(ERROR, "unsupported indexqual type: %d",
                 (int) nodeTag(clause));
            other_operand = NULL;   /* keep compiler quiet */
        }
 
        cost_qual_eval_node(&index_qual_cost, other_operand, root);
        qual_arg_cost += index_qual_cost.startup + index_qual_cost.per_tuple;
    }
    return qual_arg_cost;
}

References OpExpr::args, ScalarArrayOpExpr::args, cost_qual_eval_node(), elog, ERROR, IsA, lfirst, lsecond, nodeTag, QualCost::per_tuple, RowCompareExpr::rargs, root, and QualCost::startup.

Referenced by brincostestimate(), genericcostestimate(), and gincostestimate().

ineq_histogram_selectivity()

double ineq_histogram_selectivity	(	PlannerInfo *	root,
		VariableStatData *	vardata,
		Oid	opoid,
		FmgrInfo *	opproc,
		bool	isgt,
		bool	iseq,
		Oid	collation,
		Datum	constval,
		Oid	consttype
	)

Definition at line 1114 of file selfuncs.c.

{
    double      hist_selec;
    AttStatsSlot sslot;
 
    hist_selec = -1.0;
 
    /*
     * Someday, ANALYZE might store more than one histogram per rel/att,
     * corresponding to more than one possible sort ordering defined for the
     * column type.  Right now, we know there is only one, so just grab it and
     * see if it matches the query.
     *
     * Note that we can't use opoid as search argument; the staop appearing in
     * pg_statistic will be for the relevant '<' operator, but what we have
     * might be some other inequality operator such as '>='.  (Even if opoid
     * is a '<' operator, it could be cross-type.)  Hence we must use
     * comparison_ops_are_compatible() to see if the operators match.
     */
    if (HeapTupleIsValid(vardata->statsTuple) &&
        statistic_proc_security_check(vardata, opproc->fn_oid) &&
        get_attstatsslot(&sslot, vardata->statsTuple,
                         STATISTIC_KIND_HISTOGRAM, InvalidOid,
                         ATTSTATSSLOT_VALUES))
    {
        if (sslot.nvalues > 1 &&
            sslot.stacoll == collation &&
            comparison_ops_are_compatible(sslot.staop, opoid))
        {
            /*
             * Use binary search to find the desired location, namely the
             * right end of the histogram bin containing the comparison value,
             * which is the leftmost entry for which the comparison operator
             * succeeds (if isgt) or fails (if !isgt).
             *
             * In this loop, we pay no attention to whether the operator iseq
             * or not; that detail will be mopped up below.  (We cannot tell,
             * anyway, whether the operator thinks the values are equal.)
             *
             * If the binary search accesses the first or last histogram
             * entry, we try to replace that endpoint with the true column min
             * or max as found by get_actual_variable_range().  This
             * ameliorates misestimates when the min or max is moving as a
             * result of changes since the last ANALYZE.  Note that this could
             * result in effectively including MCVs into the histogram that
             * weren't there before, but we don't try to correct for that.
             */
            double      histfrac;
            int         lobound = 0;    /* first possible slot to search */
            int         hibound = sslot.nvalues;    /* last+1 slot to search */
            bool        have_end = false;
 
            /*
             * If there are only two histogram entries, we'll want up-to-date
             * values for both.  (If there are more than two, we need at most
             * one of them to be updated, so we deal with that within the
             * loop.)
             */
            if (sslot.nvalues == 2)
                have_end = get_actual_variable_range(root,
                                                     vardata,
                                                     sslot.staop,
                                                     collation,
                                                     &sslot.values[0],
                                                     &sslot.values[1]);
 
            while (lobound < hibound)
            {
                int         probe = (lobound + hibound) / 2;
                bool        ltcmp;
 
                /*
                 * If we find ourselves about to compare to the first or last
                 * histogram entry, first try to replace it with the actual
                 * current min or max (unless we already did so above).
                 */
                if (probe == 0 && sslot.nvalues > 2)
                    have_end = get_actual_variable_range(root,
                                                         vardata,
                                                         sslot.staop,
                                                         collation,
                                                         &sslot.values[0],
                                                         NULL);
                else if (probe == sslot.nvalues - 1 && sslot.nvalues > 2)
                    have_end = get_actual_variable_range(root,
                                                         vardata,
                                                         sslot.staop,
                                                         collation,
                                                         NULL,
                                                         &sslot.values[probe]);
 
                ltcmp = DatumGetBool(FunctionCall2Coll(opproc,
                                                       collation,
                                                       sslot.values[probe],
                                                       constval));
                if (isgt)
                    ltcmp = !ltcmp;
                if (ltcmp)
                    lobound = probe + 1;
                else
                    hibound = probe;
            }
 
            if (lobound <= 0)
            {
                /*
                 * Constant is below lower histogram boundary.  More
                 * precisely, we have found that no entry in the histogram
                 * satisfies the inequality clause (if !isgt) or they all do
                 * (if isgt).  We estimate that that's true of the entire
                 * table, so set histfrac to 0.0 (which we'll flip to 1.0
                 * below, if isgt).
                 */
                histfrac = 0.0;
            }
            else if (lobound >= sslot.nvalues)
            {
                /*
                 * Inverse case: constant is above upper histogram boundary.
                 */
                histfrac = 1.0;
            }
            else
            {
                /* We have values[i-1] <= constant <= values[i]. */
                int         i = lobound;
                double      eq_selec = 0;
                double      val,
                            high,
                            low;
                double      binfrac;
 
                /*
                 * In the cases where we'll need it below, obtain an estimate
                 * of the selectivity of "x = constval".  We use a calculation
                 * similar to what var_eq_const() does for a non-MCV constant,
                 * ie, estimate that all distinct non-MCV values occur equally
                 * often.  But multiplication by "1.0 - sumcommon - nullfrac"
                 * will be done by our caller, so we shouldn't do that here.
                 * Therefore we can't try to clamp the estimate by reference
                 * to the least common MCV; the result would be too small.
                 *
                 * Note: since this is effectively assuming that constval
                 * isn't an MCV, it's logically dubious if constval in fact is
                 * one.  But we have to apply *some* correction for equality,
                 * and anyway we cannot tell if constval is an MCV, since we
                 * don't have a suitable equality operator at hand.
                 */
                if (i == 1 || isgt == iseq)
                {
                    double      otherdistinct;
                    bool        isdefault;
                    AttStatsSlot mcvslot;
 
                    /* Get estimated number of distinct values */
                    otherdistinct = get_variable_numdistinct(vardata,
                                                             &isdefault);
 
                    /* Subtract off the number of known MCVs */
                    if (get_attstatsslot(&mcvslot, vardata->statsTuple,
                                         STATISTIC_KIND_MCV, InvalidOid,
                                         ATTSTATSSLOT_NUMBERS))
                    {
                        otherdistinct -= mcvslot.nnumbers;
                        free_attstatsslot(&mcvslot);
                    }
 
                    /* If result doesn't seem sane, leave eq_selec at 0 */
                    if (otherdistinct > 1)
                        eq_selec = 1.0 / otherdistinct;
                }
 
                /*
                 * Convert the constant and the two nearest bin boundary
                 * values to a uniform comparison scale, and do a linear
                 * interpolation within this bin.
                 */
                if (convert_to_scalar(constval, consttype, collation,
                                      &val,
                                      sslot.values[i - 1], sslot.values[i],
                                      vardata->vartype,
                                      &low, &high))
                {
                    if (high <= low)
                    {
                        /* cope if bin boundaries appear identical */
                        binfrac = 0.5;
                    }
                    else if (val <= low)
                        binfrac = 0.0;
                    else if (val >= high)
                        binfrac = 1.0;
                    else
                    {
                        binfrac = (val - low) / (high - low);
 
                        /*
                         * Watch out for the possibility that we got a NaN or
                         * Infinity from the division.  This can happen
                         * despite the previous checks, if for example "low"
                         * is -Infinity.
                         */
                        if (isnan(binfrac) ||
                            binfrac < 0.0 || binfrac > 1.0)
                            binfrac = 0.5;
                    }
                }
                else
                {
                    /*
                     * Ideally we'd produce an error here, on the grounds that
                     * the given operator shouldn't have scalarXXsel
                     * registered as its selectivity func unless we can deal
                     * with its operand types.  But currently, all manner of
                     * stuff is invoking scalarXXsel, so give a default
                     * estimate until that can be fixed.
                     */
                    binfrac = 0.5;
                }
 
                /*
                 * Now, compute the overall selectivity across the values
                 * represented by the histogram.  We have i-1 full bins and
                 * binfrac partial bin below the constant.
                 */
                histfrac = (double) (i - 1) + binfrac;
                histfrac /= (double) (sslot.nvalues - 1);
 
                /*
                 * At this point, histfrac is an estimate of the fraction of
                 * the population represented by the histogram that satisfies
                 * "x <= constval".  Somewhat remarkably, this statement is
                 * true regardless of which operator we were doing the probes
                 * with, so long as convert_to_scalar() delivers reasonable
                 * results.  If the probe constant is equal to some histogram
                 * entry, we would have considered the bin to the left of that
                 * entry if probing with "<" or ">=", or the bin to the right
                 * if probing with "<=" or ">"; but binfrac would have come
                 * out as 1.0 in the first case and 0.0 in the second, leading
                 * to the same histfrac in either case.  For probe constants
                 * between histogram entries, we find the same bin and get the
                 * same estimate with any operator.
                 *
                 * The fact that the estimate corresponds to "x <= constval"
                 * and not "x < constval" is because of the way that ANALYZE
                 * constructs the histogram: each entry is, effectively, the
                 * rightmost value in its sample bucket.  So selectivity
                 * values that are exact multiples of 1/(histogram_size-1)
                 * should be understood as estimates including a histogram
                 * entry plus everything to its left.
                 *
                 * However, that breaks down for the first histogram entry,
                 * which necessarily is the leftmost value in its sample
                 * bucket.  That means the first histogram bin is slightly
                 * narrower than the rest, by an amount equal to eq_selec.
                 * Another way to say that is that we want "x <= leftmost" to
                 * be estimated as eq_selec not zero.  So, if we're dealing
                 * with the first bin (i==1), rescale to make that true while
                 * adjusting the rest of that bin linearly.
                 */
                if (i == 1)
                    histfrac += eq_selec * (1.0 - binfrac);
 
                /*
                 * "x <= constval" is good if we want an estimate for "<=" or
                 * ">", but if we are estimating for "<" or ">=", we now need
                 * to decrease the estimate by eq_selec.
                 */
                if (isgt == iseq)
                    histfrac -= eq_selec;
            }
 
            /*
             * Now the estimate is finished for "<" and "<=" cases.  If we are
             * estimating for ">" or ">=", flip it.
             */
            hist_selec = isgt ? (1.0 - histfrac) : histfrac;
 
            /*
             * The histogram boundaries are only approximate to begin with,
             * and may well be out of date anyway.  Therefore, don't believe
             * extremely small or large selectivity estimates --- unless we
             * got actual current endpoint values from the table, in which
             * case just do the usual sanity clamp.  Somewhat arbitrarily, we
             * set the cutoff for other cases at a hundredth of the histogram
             * resolution.
             */
            if (have_end)
                CLAMP_PROBABILITY(hist_selec);
            else
            {
                double      cutoff = 0.01 / (double) (sslot.nvalues - 1);
 
                if (hist_selec < cutoff)
                    hist_selec = cutoff;
                else if (hist_selec > 1.0 - cutoff)
                    hist_selec = 1.0 - cutoff;
            }
        }
        else if (sslot.nvalues > 1)
        {
            /*
             * If we get here, we have a histogram but it's not sorted the way
             * we want.  Do a brute-force search to see how many of the
             * entries satisfy the comparison condition, and take that
             * fraction as our estimate.  (This is identical to the inner loop
             * of histogram_selectivity; maybe share code?)
             */
            LOCAL_FCINFO(fcinfo, 2);
            int         nmatch = 0;
 
            InitFunctionCallInfoData(*fcinfo, opproc, 2, collation,
                                     NULL, NULL);
            fcinfo->args[0].isnull = false;
            fcinfo->args[1].isnull = false;
            fcinfo->args[1].value = constval;
            for (int i = 0; i < sslot.nvalues; i++)
            {
                Datum       fresult;
 
                fcinfo->args[0].value = sslot.values[i];
                fcinfo->isnull = false;
                fresult = FunctionCallInvoke(fcinfo);
                if (!fcinfo->isnull && DatumGetBool(fresult))
                    nmatch++;
            }
            hist_selec = ((double) nmatch) / ((double) sslot.nvalues);
 
            /*
             * As above, clamp to a hundredth of the histogram resolution.
             * This case is surely even less trustworthy than the normal one,
             * so we shouldn't believe exact 0 or 1 selectivity.  (Maybe the
             * clamp should be more restrictive in this case?)
             */
            {
                double      cutoff = 0.01 / (double) (sslot.nvalues - 1);
 
                if (hist_selec < cutoff)
                    hist_selec = cutoff;
                else if (hist_selec > 1.0 - cutoff)
                    hist_selec = 1.0 - cutoff;
            }
        }
 
        free_attstatsslot(&sslot);
    }
 
    return hist_selec;
}

References ATTSTATSSLOT_NUMBERS, ATTSTATSSLOT_VALUES, CLAMP_PROBABILITY, comparison_ops_are_compatible(), convert_to_scalar(), DatumGetBool(), FmgrInfo::fn_oid, free_attstatsslot(), FunctionCall2Coll(), FunctionCallInvoke, get_actual_variable_range(), get_attstatsslot(), get_variable_numdistinct(), HeapTupleIsValid, i, InitFunctionCallInfoData, InvalidOid, LOCAL_FCINFO, AttStatsSlot::nnumbers, AttStatsSlot::nvalues, root, AttStatsSlot::stacoll, AttStatsSlot::staop, statistic_proc_security_check(), VariableStatData::statsTuple, val, AttStatsSlot::values, and VariableStatData::vartype.

Referenced by prefix_selectivity(), and scalarineqsel().

mcv_selectivity()

double mcv_selectivity	(	VariableStatData *	vardata,
		FmgrInfo *	opproc,
		Oid	collation,
		Datum	constval,
		bool	varonleft,
		double *	sumcommonp
	)

Definition at line 805 of file selfuncs.c.

{
    double      mcv_selec,
                sumcommon;
    AttStatsSlot sslot;
    int         i;
 
    mcv_selec = 0.0;
    sumcommon = 0.0;
 
    if (HeapTupleIsValid(vardata->statsTuple) &&
        statistic_proc_security_check(vardata, opproc->fn_oid) &&
        get_attstatsslot(&sslot, vardata->statsTuple,
                         STATISTIC_KIND_MCV, InvalidOid,
                         ATTSTATSSLOT_VALUES | ATTSTATSSLOT_NUMBERS))
    {
        LOCAL_FCINFO(fcinfo, 2);
 
        /*
         * We invoke the opproc "by hand" so that we won't fail on NULL
         * results.  Such cases won't arise for normal comparison functions,
         * but generic_restriction_selectivity could perhaps be used with
         * operators that can return NULL.  A small side benefit is to not
         * need to re-initialize the fcinfo struct from scratch each time.
         */
        InitFunctionCallInfoData(*fcinfo, opproc, 2, collation,
                                 NULL, NULL);
        fcinfo->args[0].isnull = false;
        fcinfo->args[1].isnull = false;
        /* be careful to apply operator right way 'round */
        if (varonleft)
            fcinfo->args[1].value = constval;
        else
            fcinfo->args[0].value = constval;
 
        for (i = 0; i < sslot.nvalues; i++)
        {
            Datum       fresult;
 
            if (varonleft)
                fcinfo->args[0].value = sslot.values[i];
            else
                fcinfo->args[1].value = sslot.values[i];
            fcinfo->isnull = false;
            fresult = FunctionCallInvoke(fcinfo);
            if (!fcinfo->isnull && DatumGetBool(fresult))
                mcv_selec += sslot.numbers[i];
            sumcommon += sslot.numbers[i];
        }
        free_attstatsslot(&sslot);
    }
 
    *sumcommonp = sumcommon;
    return mcv_selec;
}

References ATTSTATSSLOT_NUMBERS, ATTSTATSSLOT_VALUES, DatumGetBool(), FmgrInfo::fn_oid, free_attstatsslot(), FunctionCallInvoke, get_attstatsslot(), HeapTupleIsValid, i, InitFunctionCallInfoData, InvalidOid, LOCAL_FCINFO, AttStatsSlot::numbers, AttStatsSlot::nvalues, statistic_proc_security_check(), VariableStatData::statsTuple, and AttStatsSlot::values.

Referenced by generic_restriction_selectivity(), networksel(), patternsel_common(), and scalarineqsel().

mergejoinscansel()

void mergejoinscansel	(	PlannerInfo *	root,
		Node *	clause,
		Oid	opfamily,
		CompareType	cmptype,
		bool	nulls_first,
		Selectivity *	leftstart,
		Selectivity *	leftend,
		Selectivity *	rightstart,
		Selectivity *	rightend
	)

Definition at line 3285 of file selfuncs.c.

{
    Node       *left,
               *right;
    VariableStatData leftvar,
                rightvar;
    Oid         opmethod;
    int         op_strategy;
    Oid         op_lefttype;
    Oid         op_righttype;
    Oid         opno,
                collation,
                lsortop,
                rsortop,
                lstatop,
                rstatop,
                ltop,
                leop,
                revltop,
                revleop;
    StrategyNumber ltstrat,
                lestrat,
                gtstrat,
                gestrat;
    bool        isgt;
    Datum       leftmin,
                leftmax,
                rightmin,
                rightmax;
    double      selec;
 
    /* Set default results if we can't figure anything out. */
    /* XXX should default "start" fraction be a bit more than 0? */
    *leftstart = *rightstart = 0.0;
    *leftend = *rightend = 1.0;
 
    /* Deconstruct the merge clause */
    if (!is_opclause(clause))
        return;                 /* shouldn't happen */
    opno = ((OpExpr *) clause)->opno;
    collation = ((OpExpr *) clause)->inputcollid;
    left = get_leftop((Expr *) clause);
    right = get_rightop((Expr *) clause);
    if (!right)
        return;                 /* shouldn't happen */
 
    /* Look for stats for the inputs */
    examine_variable(root, left, 0, &leftvar);
    examine_variable(root, right, 0, &rightvar);
 
    opmethod = get_opfamily_method(opfamily);
 
    /* Extract the operator's declared left/right datatypes */
    get_op_opfamily_properties(opno, opfamily, false,
                               &op_strategy,
                               &op_lefttype,
                               &op_righttype);
    Assert(IndexAmTranslateStrategy(op_strategy, opmethod, opfamily, true) == COMPARE_EQ);
 
    /*
     * Look up the various operators we need.  If we don't find them all, it
     * probably means the opfamily is broken, but we just fail silently.
     *
     * Note: we expect that pg_statistic histograms will be sorted by the '<'
     * operator, regardless of which sort direction we are considering.
     */
    switch (cmptype)
    {
        case COMPARE_LT:
            isgt = false;
            ltstrat = IndexAmTranslateCompareType(COMPARE_LT, opmethod, opfamily, true);
            lestrat = IndexAmTranslateCompareType(COMPARE_LE, opmethod, opfamily, true);
            if (op_lefttype == op_righttype)
            {
                /* easy case */
                ltop = get_opfamily_member(opfamily,
                                           op_lefttype, op_righttype,
                                           ltstrat);
                leop = get_opfamily_member(opfamily,
                                           op_lefttype, op_righttype,
                                           lestrat);
                lsortop = ltop;
                rsortop = ltop;
                lstatop = lsortop;
                rstatop = rsortop;
                revltop = ltop;
                revleop = leop;
            }
            else
            {
                ltop = get_opfamily_member(opfamily,
                                           op_lefttype, op_righttype,
                                           ltstrat);
                leop = get_opfamily_member(opfamily,
                                           op_lefttype, op_righttype,
                                           lestrat);
                lsortop = get_opfamily_member(opfamily,
                                              op_lefttype, op_lefttype,
                                              ltstrat);
                rsortop = get_opfamily_member(opfamily,
                                              op_righttype, op_righttype,
                                              ltstrat);
                lstatop = lsortop;
                rstatop = rsortop;
                revltop = get_opfamily_member(opfamily,
                                              op_righttype, op_lefttype,
                                              ltstrat);
                revleop = get_opfamily_member(opfamily,
                                              op_righttype, op_lefttype,
                                              lestrat);
            }
            break;
        case COMPARE_GT:
            /* descending-order case */
            isgt = true;
            ltstrat = IndexAmTranslateCompareType(COMPARE_LT, opmethod, opfamily, true);
            gtstrat = IndexAmTranslateCompareType(COMPARE_GT, opmethod, opfamily, true);
            gestrat = IndexAmTranslateCompareType(COMPARE_GE, opmethod, opfamily, true);
            if (op_lefttype == op_righttype)
            {
                /* easy case */
                ltop = get_opfamily_member(opfamily,
                                           op_lefttype, op_righttype,
                                           gtstrat);
                leop = get_opfamily_member(opfamily,
                                           op_lefttype, op_righttype,
                                           gestrat);
                lsortop = ltop;
                rsortop = ltop;
                lstatop = get_opfamily_member(opfamily,
                                              op_lefttype, op_lefttype,
                                              ltstrat);
                rstatop = lstatop;
                revltop = ltop;
                revleop = leop;
            }
            else
            {
                ltop = get_opfamily_member(opfamily,
                                           op_lefttype, op_righttype,
                                           gtstrat);
                leop = get_opfamily_member(opfamily,
                                           op_lefttype, op_righttype,
                                           gestrat);
                lsortop = get_opfamily_member(opfamily,
                                              op_lefttype, op_lefttype,
                                              gtstrat);
                rsortop = get_opfamily_member(opfamily,
                                              op_righttype, op_righttype,
                                              gtstrat);
                lstatop = get_opfamily_member(opfamily,
                                              op_lefttype, op_lefttype,
                                              ltstrat);
                rstatop = get_opfamily_member(opfamily,
                                              op_righttype, op_righttype,
                                              ltstrat);
                revltop = get_opfamily_member(opfamily,
                                              op_righttype, op_lefttype,
                                              gtstrat);
                revleop = get_opfamily_member(opfamily,
                                              op_righttype, op_lefttype,
                                              gestrat);
            }
            break;
        default:
            goto fail;          /* shouldn't get here */
    }
 
    if (!OidIsValid(lsortop) ||
        !OidIsValid(rsortop) ||
        !OidIsValid(lstatop) ||
        !OidIsValid(rstatop) ||
        !OidIsValid(ltop) ||
        !OidIsValid(leop) ||
        !OidIsValid(revltop) ||
        !OidIsValid(revleop))
        goto fail;              /* insufficient info in catalogs */
 
    /* Try to get ranges of both inputs */
    if (!isgt)
    {
        if (!get_variable_range(root, &leftvar, lstatop, collation,
                                &leftmin, &leftmax))
            goto fail;          /* no range available from stats */
        if (!get_variable_range(root, &rightvar, rstatop, collation,
                                &rightmin, &rightmax))
            goto fail;          /* no range available from stats */
    }
    else
    {
        /* need to swap the max and min */
        if (!get_variable_range(root, &leftvar, lstatop, collation,
                                &leftmax, &leftmin))
            goto fail;          /* no range available from stats */
        if (!get_variable_range(root, &rightvar, rstatop, collation,
                                &rightmax, &rightmin))
            goto fail;          /* no range available from stats */
    }
 
    /*
     * Now, the fraction of the left variable that will be scanned is the
     * fraction that's <= the right-side maximum value.  But only believe
     * non-default estimates, else stick with our 1.0.
     */
    selec = scalarineqsel(root, leop, isgt, true, collation, &leftvar,
                          rightmax, op_righttype);
    if (selec != DEFAULT_INEQ_SEL)
        *leftend = selec;
 
    /* And similarly for the right variable. */
    selec = scalarineqsel(root, revleop, isgt, true, collation, &rightvar,
                          leftmax, op_lefttype);
    if (selec != DEFAULT_INEQ_SEL)
        *rightend = selec;
 
    /*
     * Only one of the two "end" fractions can really be less than 1.0;
     * believe the smaller estimate and reset the other one to exactly 1.0. If
     * we get exactly equal estimates (as can easily happen with self-joins),
     * believe neither.
     */
    if (*leftend > *rightend)
        *leftend = 1.0;
    else if (*leftend < *rightend)
        *rightend = 1.0;
    else
        *leftend = *rightend = 1.0;
 
    /*
     * Also, the fraction of the left variable that will be scanned before the
     * first join pair is found is the fraction that's < the right-side
     * minimum value.  But only believe non-default estimates, else stick with
     * our own default.
     */
    selec = scalarineqsel(root, ltop, isgt, false, collation, &leftvar,
                          rightmin, op_righttype);
    if (selec != DEFAULT_INEQ_SEL)
        *leftstart = selec;
 
    /* And similarly for the right variable. */
    selec = scalarineqsel(root, revltop, isgt, false, collation, &rightvar,
                          leftmin, op_lefttype);
    if (selec != DEFAULT_INEQ_SEL)
        *rightstart = selec;
 
    /*
     * Only one of the two "start" fractions can really be more than zero;
     * believe the larger estimate and reset the other one to exactly 0.0. If
     * we get exactly equal estimates (as can easily happen with self-joins),
     * believe neither.
     */
    if (*leftstart < *rightstart)
        *leftstart = 0.0;
    else if (*leftstart > *rightstart)
        *rightstart = 0.0;
    else
        *leftstart = *rightstart = 0.0;
 
    /*
     * If the sort order is nulls-first, we're going to have to skip over any
     * nulls too.  These would not have been counted by scalarineqsel, and we
     * can safely add in this fraction regardless of whether we believe
     * scalarineqsel's results or not.  But be sure to clamp the sum to 1.0!
     */
    if (nulls_first)
    {
        Form_pg_statistic stats;
 
        if (HeapTupleIsValid(leftvar.statsTuple))
        {
            stats = (Form_pg_statistic) GETSTRUCT(leftvar.statsTuple);
            *leftstart += stats->stanullfrac;
            CLAMP_PROBABILITY(*leftstart);
            *leftend += stats->stanullfrac;
            CLAMP_PROBABILITY(*leftend);
        }
        if (HeapTupleIsValid(rightvar.statsTuple))
        {
            stats = (Form_pg_statistic) GETSTRUCT(rightvar.statsTuple);
            *rightstart += stats->stanullfrac;
            CLAMP_PROBABILITY(*rightstart);
            *rightend += stats->stanullfrac;
            CLAMP_PROBABILITY(*rightend);
        }
    }
 
    /* Disbelieve start >= end, just in case that can happen */
    if (*leftstart >= *leftend)
    {
        *leftstart = 0.0;
        *leftend = 1.0;
    }
    if (*rightstart >= *rightend)
    {
        *rightstart = 0.0;
        *rightend = 1.0;
    }
 
fail:
    ReleaseVariableStats(leftvar);
    ReleaseVariableStats(rightvar);
}

References Assert(), CLAMP_PROBABILITY, COMPARE_EQ, COMPARE_GE, COMPARE_GT, COMPARE_LE, COMPARE_LT, DEFAULT_INEQ_SEL, examine_variable(), get_leftop(), get_op_opfamily_properties(), get_opfamily_member(), get_opfamily_method(), get_rightop(), get_variable_range(), GETSTRUCT(), HeapTupleIsValid, IndexAmTranslateCompareType(), IndexAmTranslateStrategy(), is_opclause(), OidIsValid, ReleaseVariableStats, root, scalarineqsel(), and VariableStatData::statsTuple.

Referenced by cached_scansel().

nulltestsel()

Selectivity nulltestsel	(	PlannerInfo *	root,
		NullTestType	nulltesttype,
		Node *	arg,
		int	varRelid,
		JoinType	jointype,
		SpecialJoinInfo *	sjinfo
	)

Definition at line 1782 of file selfuncs.c.

{
    VariableStatData vardata;
    double      selec;
 
    examine_variable(root, arg, varRelid, &vardata);
 
    if (HeapTupleIsValid(vardata.statsTuple))
    {
        Form_pg_statistic stats;
        double      freq_null;
 
        stats = (Form_pg_statistic) GETSTRUCT(vardata.statsTuple);
        freq_null = stats->stanullfrac;
 
        switch (nulltesttype)
        {
            case IS_NULL:
 
                /*
                 * Use freq_null directly.
                 */
                selec = freq_null;
                break;
            case IS_NOT_NULL:
 
                /*
                 * Select not unknown (not null) values. Calculate from
                 * freq_null.
                 */
                selec = 1.0 - freq_null;
                break;
            default:
                elog(ERROR, "unrecognized nulltesttype: %d",
                     (int) nulltesttype);
                return (Selectivity) 0; /* keep compiler quiet */
        }
    }
    else if (vardata.var && IsA(vardata.var, Var) &&
             ((Var *) vardata.var)->varattno < 0)
    {
        /*
         * There are no stats for system columns, but we know they are never
         * NULL.
         */
        selec = (nulltesttype == IS_NULL) ? 0.0 : 1.0;
    }
    else
    {
        /*
         * No ANALYZE stats available, so make a guess
         */
        switch (nulltesttype)
        {
            case IS_NULL:
                selec = DEFAULT_UNK_SEL;
                break;
            case IS_NOT_NULL:
                selec = DEFAULT_NOT_UNK_SEL;
                break;
            default:
                elog(ERROR, "unrecognized nulltesttype: %d",
                     (int) nulltesttype);
                return (Selectivity) 0; /* keep compiler quiet */
        }
    }
 
    ReleaseVariableStats(vardata);
 
    /* result should be in range, but make sure... */
    CLAMP_PROBABILITY(selec);
 
    return (Selectivity) selec;
}

References arg, CLAMP_PROBABILITY, DEFAULT_NOT_UNK_SEL, DEFAULT_UNK_SEL, elog, ERROR, examine_variable(), GETSTRUCT(), HeapTupleIsValid, IS_NOT_NULL, IS_NULL, IsA, ReleaseVariableStats, root, VariableStatData::statsTuple, and VariableStatData::var.

Referenced by clause_selectivity_ext(), and clauselist_selectivity_ext().

rowcomparesel()

Selectivity rowcomparesel	(	PlannerInfo *	root,
		RowCompareExpr *	clause,
		int	varRelid,
		JoinType	jointype,
		SpecialJoinInfo *	sjinfo
	)

Definition at line 2289 of file selfuncs.c.

{
    Selectivity s1;
    Oid         opno = linitial_oid(clause->opnos);
    Oid         inputcollid = linitial_oid(clause->inputcollids);
    List       *opargs;
    bool        is_join_clause;
 
    /* Build equivalent arg list for single operator */
    opargs = list_make2(linitial(clause->largs), linitial(clause->rargs));
 
    /*
     * Decide if it's a join clause.  This should match clausesel.c's
     * treat_as_join_clause(), except that we intentionally consider only the
     * leading columns and not the rest of the clause.
     */
    if (varRelid != 0)
    {
        /*
         * Caller is forcing restriction mode (eg, because we are examining an
         * inner indexscan qual).
         */
        is_join_clause = false;
    }
    else if (sjinfo == NULL)
    {
        /*
         * It must be a restriction clause, since it's being evaluated at a
         * scan node.
         */
        is_join_clause = false;
    }
    else
    {
        /*
         * Otherwise, it's a join if there's more than one base relation used.
         */
        is_join_clause = (NumRelids(root, (Node *) opargs) > 1);
    }
 
    if (is_join_clause)
    {
        /* Estimate selectivity for a join clause. */
        s1 = join_selectivity(root, opno,
                              opargs,
                              inputcollid,
                              jointype,
                              sjinfo);
    }
    else
    {
        /* Estimate selectivity for a restriction clause. */
        s1 = restriction_selectivity(root, opno,
                                     opargs,
                                     inputcollid,
                                     varRelid);
    }
 
    return s1;
}

References join_selectivity(), RowCompareExpr::largs, linitial, linitial_oid, list_make2, NumRelids(), RowCompareExpr::rargs, restriction_selectivity(), root, and s1.

Referenced by clause_selectivity_ext().

scalararraysel()

Selectivity scalararraysel	(	PlannerInfo *	root,
		ScalarArrayOpExpr *	clause,
		bool	is_join_clause,
		int	varRelid,
		JoinType	jointype,
		SpecialJoinInfo *	sjinfo
	)

Definition at line 1900 of file selfuncs.c.

{
    Oid         operator = clause->opno;
    bool        useOr = clause->useOr;
    bool        isEquality = false;
    bool        isInequality = false;
    Node       *leftop;
    Node       *rightop;
    Oid         nominal_element_type;
    Oid         nominal_element_collation;
    TypeCacheEntry *typentry;
    RegProcedure oprsel;
    FmgrInfo    oprselproc;
    Selectivity s1;
    Selectivity s1disjoint;
 
    /* First, deconstruct the expression */
    Assert(list_length(clause->args) == 2);
    leftop = (Node *) linitial(clause->args);
    rightop = (Node *) lsecond(clause->args);
 
    /* aggressively reduce both sides to constants */
    leftop = estimate_expression_value(root, leftop);
    rightop = estimate_expression_value(root, rightop);
 
    /* get nominal (after relabeling) element type of rightop */
    nominal_element_type = get_base_element_type(exprType(rightop));
    if (!OidIsValid(nominal_element_type))
        return (Selectivity) 0.5;   /* probably shouldn't happen */
    /* get nominal collation, too, for generating constants */
    nominal_element_collation = exprCollation(rightop);
 
    /* look through any binary-compatible relabeling of rightop */
    rightop = strip_array_coercion(rightop);
 
    /*
     * Detect whether the operator is the default equality or inequality
     * operator of the array element type.
     */
    typentry = lookup_type_cache(nominal_element_type, TYPECACHE_EQ_OPR);
    if (OidIsValid(typentry->eq_opr))
    {
        if (operator == typentry->eq_opr)
            isEquality = true;
        else if (get_negator(operator) == typentry->eq_opr)
            isInequality = true;
    }
 
    /*
     * If it is equality or inequality, we might be able to estimate this as a
     * form of array containment; for instance "const = ANY(column)" can be
     * treated as "ARRAY[const] <@ column".  scalararraysel_containment tries
     * that, and returns the selectivity estimate if successful, or -1 if not.
     */
    if ((isEquality || isInequality) && !is_join_clause)
    {
        s1 = scalararraysel_containment(root, leftop, rightop,
                                        nominal_element_type,
                                        isEquality, useOr, varRelid);
        if (s1 >= 0.0)
            return s1;
    }
 
    /*
     * Look up the underlying operator's selectivity estimator. Punt if it
     * hasn't got one.
     */
    if (is_join_clause)
        oprsel = get_oprjoin(operator);
    else
        oprsel = get_oprrest(operator);
    if (!oprsel)
        return (Selectivity) 0.5;
    fmgr_info(oprsel, &oprselproc);
 
    /*
     * In the array-containment check above, we must only believe that an
     * operator is equality or inequality if it is the default btree equality
     * operator (or its negator) for the element type, since those are the
     * operators that array containment will use.  But in what follows, we can
     * be a little laxer, and also believe that any operators using eqsel() or
     * neqsel() as selectivity estimator act like equality or inequality.
     */
    if (oprsel == F_EQSEL || oprsel == F_EQJOINSEL)
        isEquality = true;
    else if (oprsel == F_NEQSEL || oprsel == F_NEQJOINSEL)
        isInequality = true;
 
    /*
     * We consider three cases:
     *
     * 1. rightop is an Array constant: deconstruct the array, apply the
     * operator's selectivity function for each array element, and merge the
     * results in the same way that clausesel.c does for AND/OR combinations.
     *
     * 2. rightop is an ARRAY[] construct: apply the operator's selectivity
     * function for each element of the ARRAY[] construct, and merge.
     *
     * 3. otherwise, make a guess ...
     */
    if (rightop && IsA(rightop, Const))
    {
        Datum       arraydatum = ((Const *) rightop)->constvalue;
        bool        arrayisnull = ((Const *) rightop)->constisnull;
        ArrayType  *arrayval;
        int16       elmlen;
        bool        elmbyval;
        char        elmalign;
        int         num_elems;
        Datum      *elem_values;
        bool       *elem_nulls;
        int         i;
 
        if (arrayisnull)        /* qual can't succeed if null array */
            return (Selectivity) 0.0;
        arrayval = DatumGetArrayTypeP(arraydatum);
        get_typlenbyvalalign(ARR_ELEMTYPE(arrayval),
                             &elmlen, &elmbyval, &elmalign);
        deconstruct_array(arrayval,
                          ARR_ELEMTYPE(arrayval),
                          elmlen, elmbyval, elmalign,
                          &elem_values, &elem_nulls, &num_elems);
 
        /*
         * For generic operators, we assume the probability of success is
         * independent for each array element.  But for "= ANY" or "<> ALL",
         * if the array elements are distinct (which'd typically be the case)
         * then the probabilities are disjoint, and we should just sum them.
         *
         * If we were being really tense we would try to confirm that the
         * elements are all distinct, but that would be expensive and it
         * doesn't seem to be worth the cycles; it would amount to penalizing
         * well-written queries in favor of poorly-written ones.  However, we
         * do protect ourselves a little bit by checking whether the
         * disjointness assumption leads to an impossible (out of range)
         * probability; if so, we fall back to the normal calculation.
         */
        s1 = s1disjoint = (useOr ? 0.0 : 1.0);
 
        for (i = 0; i < num_elems; i++)
        {
            List       *args;
            Selectivity s2;
 
            args = list_make2(leftop,
                              makeConst(nominal_element_type,
                                        -1,
                                        nominal_element_collation,
                                        elmlen,
                                        elem_values[i],
                                        elem_nulls[i],
                                        elmbyval));
            if (is_join_clause)
                s2 = DatumGetFloat8(FunctionCall5Coll(&oprselproc,
                                                      clause->inputcollid,
                                                      PointerGetDatum(root),
                                                      ObjectIdGetDatum(operator),
                                                      PointerGetDatum(args),
                                                      Int16GetDatum(jointype),
                                                      PointerGetDatum(sjinfo)));
            else
                s2 = DatumGetFloat8(FunctionCall4Coll(&oprselproc,
                                                      clause->inputcollid,
                                                      PointerGetDatum(root),
                                                      ObjectIdGetDatum(operator),
                                                      PointerGetDatum(args),
                                                      Int32GetDatum(varRelid)));
 
            if (useOr)
            {
                s1 = s1 + s2 - s1 * s2;
                if (isEquality)
                    s1disjoint += s2;
            }
            else
            {
                s1 = s1 * s2;
                if (isInequality)
                    s1disjoint += s2 - 1.0;
            }
        }
 
        /* accept disjoint-probability estimate if in range */
        if ((useOr ? isEquality : isInequality) &&
            s1disjoint >= 0.0 && s1disjoint <= 1.0)
            s1 = s1disjoint;
    }
    else if (rightop && IsA(rightop, ArrayExpr) &&
             !((ArrayExpr *) rightop)->multidims)
    {
        ArrayExpr  *arrayexpr = (ArrayExpr *) rightop;
        int16       elmlen;
        bool        elmbyval;
        ListCell   *l;
 
        get_typlenbyval(arrayexpr->element_typeid,
                        &elmlen, &elmbyval);
 
        /*
         * We use the assumption of disjoint probabilities here too, although
         * the odds of equal array elements are rather higher if the elements
         * are not all constants (which they won't be, else constant folding
         * would have reduced the ArrayExpr to a Const).  In this path it's
         * critical to have the sanity check on the s1disjoint estimate.
         */
        s1 = s1disjoint = (useOr ? 0.0 : 1.0);
 
        foreach(l, arrayexpr->elements)
        {
            Node       *elem = (Node *) lfirst(l);
            List       *args;
            Selectivity s2;
 
            /*
             * Theoretically, if elem isn't of nominal_element_type we should
             * insert a RelabelType, but it seems unlikely that any operator
             * estimation function would really care ...
             */
            args = list_make2(leftop, elem);
            if (is_join_clause)
                s2 = DatumGetFloat8(FunctionCall5Coll(&oprselproc,
                                                      clause->inputcollid,
                                                      PointerGetDatum(root),
                                                      ObjectIdGetDatum(operator),
                                                      PointerGetDatum(args),
                                                      Int16GetDatum(jointype),
                                                      PointerGetDatum(sjinfo)));
            else
                s2 = DatumGetFloat8(FunctionCall4Coll(&oprselproc,
                                                      clause->inputcollid,
                                                      PointerGetDatum(root),
                                                      ObjectIdGetDatum(operator),
                                                      PointerGetDatum(args),
                                                      Int32GetDatum(varRelid)));
 
            if (useOr)
            {
                s1 = s1 + s2 - s1 * s2;
                if (isEquality)
                    s1disjoint += s2;
            }
            else
            {
                s1 = s1 * s2;
                if (isInequality)
                    s1disjoint += s2 - 1.0;
            }
        }
 
        /* accept disjoint-probability estimate if in range */
        if ((useOr ? isEquality : isInequality) &&
            s1disjoint >= 0.0 && s1disjoint <= 1.0)
            s1 = s1disjoint;
    }
    else
    {
        CaseTestExpr *dummyexpr;
        List       *args;
        Selectivity s2;
        int         i;
 
        /*
         * We need a dummy rightop to pass to the operator selectivity
         * routine.  It can be pretty much anything that doesn't look like a
         * constant; CaseTestExpr is a convenient choice.
         */
        dummyexpr = makeNode(CaseTestExpr);
        dummyexpr->typeId = nominal_element_type;
        dummyexpr->typeMod = -1;
        dummyexpr->collation = clause->inputcollid;
        args = list_make2(leftop, dummyexpr);
        if (is_join_clause)
            s2 = DatumGetFloat8(FunctionCall5Coll(&oprselproc,
                                                  clause->inputcollid,
                                                  PointerGetDatum(root),
                                                  ObjectIdGetDatum(operator),
                                                  PointerGetDatum(args),
                                                  Int16GetDatum(jointype),
                                                  PointerGetDatum(sjinfo)));
        else
            s2 = DatumGetFloat8(FunctionCall4Coll(&oprselproc,
                                                  clause->inputcollid,
                                                  PointerGetDatum(root),
                                                  ObjectIdGetDatum(operator),
                                                  PointerGetDatum(args),
                                                  Int32GetDatum(varRelid)));
        s1 = useOr ? 0.0 : 1.0;
 
        /*
         * Arbitrarily assume 10 elements in the eventual array value (see
         * also estimate_array_length).  We don't risk an assumption of
         * disjoint probabilities here.
         */
        for (i = 0; i < 10; i++)
        {
            if (useOr)
                s1 = s1 + s2 - s1 * s2;
            else
                s1 = s1 * s2;
        }
    }
 
    /* result should be in range, but make sure... */
    CLAMP_PROBABILITY(s1);
 
    return s1;
}

References generate_unaccent_rules::args, ScalarArrayOpExpr::args, ARR_ELEMTYPE, Assert(), CLAMP_PROBABILITY, DatumGetArrayTypeP, DatumGetFloat8(), deconstruct_array(), TypeCacheEntry::eq_opr, estimate_expression_value(), exprCollation(), exprType(), fmgr_info(), FunctionCall4Coll(), FunctionCall5Coll(), get_base_element_type(), get_negator(), get_oprjoin(), get_oprrest(), get_typlenbyval(), get_typlenbyvalalign(), i, Int16GetDatum(), Int32GetDatum(), IsA, lfirst, linitial, list_length(), list_make2, lookup_type_cache(), lsecond, makeConst(), makeNode, ObjectIdGetDatum(), OidIsValid, ScalarArrayOpExpr::opno, PointerGetDatum(), root, s1, s2, scalararraysel_containment(), strip_array_coercion(), TYPECACHE_EQ_OPR, CaseTestExpr::typeId, and ScalarArrayOpExpr::useOr.

Referenced by clause_selectivity_ext().

scalararraysel_containment()

Selectivity scalararraysel_containment	(	PlannerInfo *	root,
		Node *	leftop,
		Node *	rightop,
		Oid	elemtype,
		bool	isEquality,
		bool	useOr,
		int	varRelid
	)

Definition at line 81 of file array_selfuncs.c.

{
    Selectivity selec;
    VariableStatData vardata;
    Datum       constval;
    TypeCacheEntry *typentry;
    FmgrInfo   *cmpfunc;
 
    /*
     * rightop must be a variable, else punt.
     */
    examine_variable(root, rightop, varRelid, &vardata);
    if (!vardata.rel)
    {
        ReleaseVariableStats(vardata);
        return -1.0;
    }
 
    /*
     * leftop must be a constant, else punt.
     */
    if (!IsA(leftop, Const))
    {
        ReleaseVariableStats(vardata);
        return -1.0;
    }
    if (((Const *) leftop)->constisnull)
    {
        /* qual can't succeed if null on left */
        ReleaseVariableStats(vardata);
        return (Selectivity) 0.0;
    }
    constval = ((Const *) leftop)->constvalue;
 
    /* Get element type's default comparison function */
    typentry = lookup_type_cache(elemtype, TYPECACHE_CMP_PROC_FINFO);
    if (!OidIsValid(typentry->cmp_proc_finfo.fn_oid))
    {
        ReleaseVariableStats(vardata);
        return -1.0;
    }
    cmpfunc = &typentry->cmp_proc_finfo;
 
    /*
     * If the operator is <>, swap ANY/ALL, then invert the result later.
     */
    if (!isEquality)
        useOr = !useOr;
 
    /* Get array element stats for var, if available */
    if (HeapTupleIsValid(vardata.statsTuple) &&
        statistic_proc_security_check(&vardata, cmpfunc->fn_oid))
    {
        Form_pg_statistic stats;
        AttStatsSlot sslot;
        AttStatsSlot hslot;
 
        stats = (Form_pg_statistic) GETSTRUCT(vardata.statsTuple);
 
        /* MCELEM will be an array of same type as element */
        if (get_attstatsslot(&sslot, vardata.statsTuple,
                             STATISTIC_KIND_MCELEM, InvalidOid,
                             ATTSTATSSLOT_VALUES | ATTSTATSSLOT_NUMBERS))
        {
            /* For ALL case, also get histogram of distinct-element counts */
            if (useOr ||
                !get_attstatsslot(&hslot, vardata.statsTuple,
                                  STATISTIC_KIND_DECHIST, InvalidOid,
                                  ATTSTATSSLOT_NUMBERS))
                memset(&hslot, 0, sizeof(hslot));
 
            /*
             * For = ANY, estimate as var @> ARRAY[const].
             *
             * For = ALL, estimate as var <@ ARRAY[const].
             */
            if (useOr)
                selec = mcelem_array_contain_overlap_selec(sslot.values,
                                                           sslot.nvalues,
                                                           sslot.numbers,
                                                           sslot.nnumbers,
                                                           &constval, 1,
                                                           OID_ARRAY_CONTAINS_OP,
                                                           typentry);
            else
                selec = mcelem_array_contained_selec(sslot.values,
                                                     sslot.nvalues,
                                                     sslot.numbers,
                                                     sslot.nnumbers,
                                                     &constval, 1,
                                                     hslot.numbers,
                                                     hslot.nnumbers,
                                                     OID_ARRAY_CONTAINED_OP,
                                                     typentry);
 
            free_attstatsslot(&hslot);
            free_attstatsslot(&sslot);
        }
        else
        {
            /* No most-common-elements info, so do without */
            if (useOr)
                selec = mcelem_array_contain_overlap_selec(NULL, 0,
                                                           NULL, 0,
                                                           &constval, 1,
                                                           OID_ARRAY_CONTAINS_OP,
                                                           typentry);
            else
                selec = mcelem_array_contained_selec(NULL, 0,
                                                     NULL, 0,
                                                     &constval, 1,
                                                     NULL, 0,
                                                     OID_ARRAY_CONTAINED_OP,
                                                     typentry);
        }
 
        /*
         * MCE stats count only non-null rows, so adjust for null rows.
         */
        selec *= (1.0 - stats->stanullfrac);
    }
    else
    {
        /* No stats at all, so do without */
        if (useOr)
            selec = mcelem_array_contain_overlap_selec(NULL, 0,
                                                       NULL, 0,
                                                       &constval, 1,
                                                       OID_ARRAY_CONTAINS_OP,
                                                       typentry);
        else
            selec = mcelem_array_contained_selec(NULL, 0,
                                                 NULL, 0,
                                                 &constval, 1,
                                                 NULL, 0,
                                                 OID_ARRAY_CONTAINED_OP,
                                                 typentry);
        /* we assume no nulls here, so no stanullfrac correction */
    }
 
    ReleaseVariableStats(vardata);
 
    /*
     * If the operator is <>, invert the results.
     */
    if (!isEquality)
        selec = 1.0 - selec;
 
    CLAMP_PROBABILITY(selec);
 
    return selec;
}

References ATTSTATSSLOT_NUMBERS, ATTSTATSSLOT_VALUES, CLAMP_PROBABILITY, TypeCacheEntry::cmp_proc_finfo, examine_variable(), FmgrInfo::fn_oid, free_attstatsslot(), get_attstatsslot(), GETSTRUCT(), HeapTupleIsValid, InvalidOid, IsA, lookup_type_cache(), mcelem_array_contain_overlap_selec(), mcelem_array_contained_selec(), AttStatsSlot::nnumbers, AttStatsSlot::numbers, AttStatsSlot::nvalues, OidIsValid, VariableStatData::rel, ReleaseVariableStats, root, statistic_proc_security_check(), VariableStatData::statsTuple, TYPECACHE_CMP_PROC_FINFO, and AttStatsSlot::values.

Referenced by scalararraysel().

statistic_proc_security_check()

bool statistic_proc_security_check	(	VariableStatData *	vardata,
		Oid	func_oid
	)

Definition at line 6573 of file selfuncs.c.

{
    if (vardata->acl_ok)
        return true;            /* have SELECT privs and no securityQuals */
 
    if (!OidIsValid(func_oid))
        return false;
 
    if (get_func_leakproof(func_oid))
        return true;
 
    ereport(DEBUG2,
            (errmsg_internal("not using statistics because function \"%s\" is not leakproof",
                             get_func_name(func_oid))));
    return false;
}

References VariableStatData::acl_ok, DEBUG2, ereport, errmsg_internal(), get_func_leakproof(), get_func_name(), and OidIsValid.

Referenced by calc_arraycontsel(), calc_hist_selectivity(), eqjoinsel(), get_variable_range(), histogram_selectivity(), ineq_histogram_selectivity(), mcv_selectivity(), scalararraysel_containment(), and var_eq_const().

var_eq_const()

double var_eq_const	(	VariableStatData *	vardata,
		Oid	oproid,
		Oid	collation,
		Datum	constval,
		bool	constisnull,
		bool	varonleft,
		bool	negate
	)

Definition at line 368 of file selfuncs.c.

{
    double      selec;
    double      nullfrac = 0.0;
    bool        isdefault;
    Oid         opfuncoid;
 
    /*
     * If the constant is NULL, assume operator is strict and return zero, ie,
     * operator will never return TRUE.  (It's zero even for a negator op.)
     */
    if (constisnull)
        return 0.0;
 
    /*
     * Grab the nullfrac for use below.  Note we allow use of nullfrac
     * regardless of security check.
     */
    if (HeapTupleIsValid(vardata->statsTuple))
    {
        Form_pg_statistic stats;
 
        stats = (Form_pg_statistic) GETSTRUCT(vardata->statsTuple);
        nullfrac = stats->stanullfrac;
    }
 
    /*
     * If we matched the var to a unique index, DISTINCT or GROUP-BY clause,
     * assume there is exactly one match regardless of anything else.  (This
     * is slightly bogus, since the index or clause's equality operator might
     * be different from ours, but it's much more likely to be right than
     * ignoring the information.)
     */
    if (vardata->isunique && vardata->rel && vardata->rel->tuples >= 1.0)
    {
        selec = 1.0 / vardata->rel->tuples;
    }
    else if (HeapTupleIsValid(vardata->statsTuple) &&
             statistic_proc_security_check(vardata,
                                           (opfuncoid = get_opcode(oproid))))
    {
        AttStatsSlot sslot;
        bool        match = false;
        int         i;
 
        /*
         * Is the constant "=" to any of the column's most common values?
         * (Although the given operator may not really be "=", we will assume
         * that seeing whether it returns TRUE is an appropriate test.  If you
         * don't like this, maybe you shouldn't be using eqsel for your
         * operator...)
         */
        if (get_attstatsslot(&sslot, vardata->statsTuple,
                             STATISTIC_KIND_MCV, InvalidOid,
                             ATTSTATSSLOT_VALUES | ATTSTATSSLOT_NUMBERS))
        {
            LOCAL_FCINFO(fcinfo, 2);
            FmgrInfo    eqproc;
 
            fmgr_info(opfuncoid, &eqproc);
 
            /*
             * Save a few cycles by setting up the fcinfo struct just once.
             * Using FunctionCallInvoke directly also avoids failure if the
             * eqproc returns NULL, though really equality functions should
             * never do that.
             */
            InitFunctionCallInfoData(*fcinfo, &eqproc, 2, collation,
                                     NULL, NULL);
            fcinfo->args[0].isnull = false;
            fcinfo->args[1].isnull = false;
            /* be careful to apply operator right way 'round */
            if (varonleft)
                fcinfo->args[1].value = constval;
            else
                fcinfo->args[0].value = constval;
 
            for (i = 0; i < sslot.nvalues; i++)
            {
                Datum       fresult;
 
                if (varonleft)
                    fcinfo->args[0].value = sslot.values[i];
                else
                    fcinfo->args[1].value = sslot.values[i];
                fcinfo->isnull = false;
                fresult = FunctionCallInvoke(fcinfo);
                if (!fcinfo->isnull && DatumGetBool(fresult))
                {
                    match = true;
                    break;
                }
            }
        }
        else
        {
            /* no most-common-value info available */
            i = 0;              /* keep compiler quiet */
        }
 
        if (match)
        {
            /*
             * Constant is "=" to this common value.  We know selectivity
             * exactly (or as exactly as ANALYZE could calculate it, anyway).
             */
            selec = sslot.numbers[i];
        }
        else
        {
            /*
             * Comparison is against a constant that is neither NULL nor any
             * of the common values.  Its selectivity cannot be more than
             * this:
             */
            double      sumcommon = 0.0;
            double      otherdistinct;
 
            for (i = 0; i < sslot.nnumbers; i++)
                sumcommon += sslot.numbers[i];
            selec = 1.0 - sumcommon - nullfrac;
            CLAMP_PROBABILITY(selec);
 
            /*
             * and in fact it's probably a good deal less. We approximate that
             * all the not-common values share this remaining fraction
             * equally, so we divide by the number of other distinct values.
             */
            otherdistinct = get_variable_numdistinct(vardata, &isdefault) -
                sslot.nnumbers;
            if (otherdistinct > 1)
                selec /= otherdistinct;
 
            /*
             * Another cross-check: selectivity shouldn't be estimated as more
             * than the least common "most common value".
             */
            if (sslot.nnumbers > 0 && selec > sslot.numbers[sslot.nnumbers - 1])
                selec = sslot.numbers[sslot.nnumbers - 1];
        }
 
        free_attstatsslot(&sslot);
    }
    else
    {
        /*
         * No ANALYZE stats available, so make a guess using estimated number
         * of distinct values and assuming they are equally common. (The guess
         * is unlikely to be very good, but we do know a few special cases.)
         */
        selec = 1.0 / get_variable_numdistinct(vardata, &isdefault);
    }
 
    /* now adjust if we wanted <> rather than = */
    if (negate)
        selec = 1.0 - selec - nullfrac;
 
    /* result should be in range, but make sure... */
    CLAMP_PROBABILITY(selec);
 
    return selec;
}

References ATTSTATSSLOT_NUMBERS, ATTSTATSSLOT_VALUES, CLAMP_PROBABILITY, DatumGetBool(), fmgr_info(), free_attstatsslot(), FunctionCallInvoke, get_attstatsslot(), get_opcode(), get_variable_numdistinct(), GETSTRUCT(), HeapTupleIsValid, i, InitFunctionCallInfoData, InvalidOid, VariableStatData::isunique, LOCAL_FCINFO, AttStatsSlot::nnumbers, AttStatsSlot::numbers, AttStatsSlot::nvalues, VariableStatData::rel, statistic_proc_security_check(), VariableStatData::statsTuple, RelOptInfo::tuples, and AttStatsSlot::values.

Referenced by boolvarsel(), eqsel_internal(), patternsel_common(), and prefix_selectivity().

var_eq_non_const()

double var_eq_non_const	(	VariableStatData *	vardata,
		Oid	oproid,
		Oid	collation,
		Node *	other,
		bool	varonleft,
		bool	negate
	)

Definition at line 539 of file selfuncs.c.

{
    double      selec;
    double      nullfrac = 0.0;
    bool        isdefault;
 
    /*
     * Grab the nullfrac for use below.
     */
    if (HeapTupleIsValid(vardata->statsTuple))
    {
        Form_pg_statistic stats;
 
        stats = (Form_pg_statistic) GETSTRUCT(vardata->statsTuple);
        nullfrac = stats->stanullfrac;
    }
 
    /*
     * If we matched the var to a unique index, DISTINCT or GROUP-BY clause,
     * assume there is exactly one match regardless of anything else.  (This
     * is slightly bogus, since the index or clause's equality operator might
     * be different from ours, but it's much more likely to be right than
     * ignoring the information.)
     */
    if (vardata->isunique && vardata->rel && vardata->rel->tuples >= 1.0)
    {
        selec = 1.0 / vardata->rel->tuples;
    }
    else if (HeapTupleIsValid(vardata->statsTuple))
    {
        double      ndistinct;
        AttStatsSlot sslot;
 
        /*
         * Search is for a value that we do not know a priori, but we will
         * assume it is not NULL.  Estimate the selectivity as non-null
         * fraction divided by number of distinct values, so that we get a
         * result averaged over all possible values whether common or
         * uncommon.  (Essentially, we are assuming that the not-yet-known
         * comparison value is equally likely to be any of the possible
         * values, regardless of their frequency in the table.  Is that a good
         * idea?)
         */
        selec = 1.0 - nullfrac;
        ndistinct = get_variable_numdistinct(vardata, &isdefault);
        if (ndistinct > 1)
            selec /= ndistinct;
 
        /*
         * Cross-check: selectivity should never be estimated as more than the
         * most common value's.
         */
        if (get_attstatsslot(&sslot, vardata->statsTuple,
                             STATISTIC_KIND_MCV, InvalidOid,
                             ATTSTATSSLOT_NUMBERS))
        {
            if (sslot.nnumbers > 0 && selec > sslot.numbers[0])
                selec = sslot.numbers[0];
            free_attstatsslot(&sslot);
        }
    }
    else
    {
        /*
         * No ANALYZE stats available, so make a guess using estimated number
         * of distinct values and assuming they are equally common. (The guess
         * is unlikely to be very good, but we do know a few special cases.)
         */
        selec = 1.0 / get_variable_numdistinct(vardata, &isdefault);
    }
 
    /* now adjust if we wanted <> rather than = */
    if (negate)
        selec = 1.0 - selec - nullfrac;
 
    /* result should be in range, but make sure... */
    CLAMP_PROBABILITY(selec);
 
    return selec;
}

References ATTSTATSSLOT_NUMBERS, CLAMP_PROBABILITY, free_attstatsslot(), get_attstatsslot(), get_variable_numdistinct(), GETSTRUCT(), HeapTupleIsValid, InvalidOid, VariableStatData::isunique, AttStatsSlot::nnumbers, AttStatsSlot::numbers, VariableStatData::rel, VariableStatData::statsTuple, and RelOptInfo::tuples.

Referenced by eqsel_internal().

Variable Documentation

get_index_stats_hook

PGDLLIMPORT get_index_stats_hook_type get_index_stats_hook

extern

Definition at line 184 of file selfuncs.c.

Referenced by brincostestimate(), examine_indexcol_variable(), and examine_variable().

get_relation_stats_hook

PGDLLIMPORT get_relation_stats_hook_type get_relation_stats_hook

extern

Definition at line 183 of file selfuncs.c.

Referenced by brincostestimate(), examine_indexcol_variable(), and examine_simple_variable().