relation.h

PostgreSQL7.4.6 for Linux

 * join clauses when plans are built.
 *
 * 'indexscandir' is one of:
 *		ForwardScanDirection: forward scan of an ordered index
 *		BackwardScanDirection: backward scan of an ordered index
 *		NoMovementScanDirection: scan of an unordered index, or don't care
 * (The executor doesn't care whether it gets ForwardScanDirection or
 * NoMovementScanDirection for an indexscan, but the planner wants to
 * distinguish ordered from unordered indexes for building pathkeys.)
 *
 * 'rows' is the estimated result tuple count for the indexscan.  This
 * is the same as path.parent->rows for a simple indexscan, but it is
 * different for a nestloop inner scan, because the additional indexquals
 * coming from join clauses make the scan more selective than the parent
 * rel's restrict clauses alone would do.
 *----------
 */
typedef struct IndexPath
{
	Path		path;
	List	   *indexinfo;
	List	   *indexqual;
	List	   *indexjoinclauses;
	ScanDirection indexscandir;
	double		rows;			/* estimated number of result tuples */
} IndexPath;

/*
 * TidPath represents a scan by TID
 */
typedef struct TidPath
{
	Path		path;
	List	   *tideval;		/* qual(s) involving CTID = something */
} TidPath;

/*
 * AppendPath represents an Append plan, ie, successive execution of
 * several member plans.  Currently it is only used to handle expansion
 * of inheritance trees.
 */
typedef struct AppendPath
{
	Path		path;
	List	   *subpaths;		/* list of component Paths */
} AppendPath;

/*
 * ResultPath represents use of a Result plan node, either to compute a
 * variable-free targetlist or to gate execution of a subplan with a
 * one-time (variable-free) qual condition.  Note that in the former case
 * path.parent will be NULL; in the latter case it is copied from the subpath.
 */
typedef struct ResultPath
{
	Path		path;
	Path	   *subpath;
	List	   *constantqual;
} ResultPath;

/*
 * MaterialPath represents use of a Material plan node, i.e., caching of
 * the output of its subpath.  This is used when the subpath is expensive
 * and needs to be scanned repeatedly, or when we need mark/restore ability
 * and the subpath doesn't have it.
 */
typedef struct MaterialPath
{
	Path		path;
	Path	   *subpath;
} MaterialPath;

/*
 * UniquePath represents elimination of distinct rows from the output of
 * its subpath.
 *
 * This is unlike the other Path nodes in that it can actually generate
 * two different plans: either hash-based or sort-based implementation.
 * The decision is sufficiently localized that it's not worth having two
 * separate Path node types.
 */
typedef struct UniquePath
{
	Path		path;
	Path	   *subpath;
	bool		use_hash;
	double		rows;			/* estimated number of result tuples */
} UniquePath;

/*
 * All join-type paths share these fields.
 */

typedef struct JoinPath
{
	Path		path;

	JoinType	jointype;

	Path	   *outerjoinpath;	/* path for the outer side of the join */
	Path	   *innerjoinpath;	/* path for the inner side of the join */

	List	   *joinrestrictinfo;		/* RestrictInfos to apply to join */

	/*
	 * See the notes for RelOptInfo to understand why joinrestrictinfo is
	 * needed in JoinPath, and can't be merged into the parent RelOptInfo.
	 */
} JoinPath;

/*
 * A nested-loop path needs no special fields.
 */

typedef JoinPath NestPath;

/*
 * A mergejoin path has these fields.
 *
 * path_mergeclauses lists the clauses (in the form of RestrictInfos)
 * that will be used in the merge.	(Before 7.0, this was a list of bare
 * clause expressions, but we can save on list memory and cost_qual_eval
 * work by leaving it in the form of a RestrictInfo list.)
 *
 * Note that the mergeclauses are a subset of the parent relation's
 * restriction-clause list.  Any join clauses that are not mergejoinable
 * appear only in the parent's restrict list, and must be checked by a
 * qpqual at execution time.
 *
 * outersortkeys (resp. innersortkeys) is NIL if the outer path
 * (resp. inner path) is already ordered appropriately for the
 * mergejoin.  If it is not NIL then it is a PathKeys list describing
 * the ordering that must be created by an explicit sort step.
 */

typedef struct MergePath
{
	JoinPath	jpath;
	List	   *path_mergeclauses;		/* join clauses to be used for
										 * merge */
	List	   *outersortkeys;	/* keys for explicit sort, if any */
	List	   *innersortkeys;	/* keys for explicit sort, if any */
} MergePath;

/*
 * A hashjoin path has these fields.
 *
 * The remarks above for mergeclauses apply for hashclauses as well.
 *
 * Hashjoin does not care what order its inputs appear in, so we have
 * no need for sortkeys.
 */

typedef struct HashPath
{
	JoinPath	jpath;
	List	   *path_hashclauses;		/* join clauses used for hashing */
} HashPath;

/*
 * Restriction clause info.
 *
 * We create one of these for each AND sub-clause of a restriction condition
 * (WHERE or JOIN/ON clause).  Since the restriction clauses are logically
 * ANDed, we can use any one of them or any subset of them to filter out
 * tuples, without having to evaluate the rest.  The RestrictInfo node itself
 * stores data used by the optimizer while choosing the best query plan.
 *
 * If a restriction clause references a single base relation, it will appear
 * in the baserestrictinfo list of the RelOptInfo for that base rel.
 *
 * If a restriction clause references more than one base rel, it will
 * appear in the JoinInfo lists of every RelOptInfo that describes a strict
 * subset of the base rels mentioned in the clause.  The JoinInfo lists are
 * used to drive join tree building by selecting plausible join candidates.
 * The clause cannot actually be applied until we have built a join rel
 * containing all the base rels it references, however.
 *
 * When we construct a join rel that includes all the base rels referenced
 * in a multi-relation restriction clause, we place that clause into the
 * joinrestrictinfo lists of paths for the join rel, if neither left nor
 * right sub-path includes all base rels referenced in the clause.	The clause
 * will be applied at that join level, and will not propagate any further up
 * the join tree.  (Note: the "predicate migration" code was once intended to
 * push restriction clauses up and down the plan tree based on evaluation
 * costs, but it's dead code and is unlikely to be resurrected in the
 * foreseeable future.)
 *
 * Note that in the presence of more than two rels, a multi-rel restriction
 * might reach different heights in the join tree depending on the join
 * sequence we use.  So, these clauses cannot be associated directly with
 * the join RelOptInfo, but must be kept track of on a per-join-path basis.
 *
 * When dealing with outer joins we have to be very careful about pushing qual
 * clauses up and down the tree.  An outer join's own JOIN/ON conditions must
 * be evaluated exactly at that join node, and any quals appearing in WHERE or
 * in a JOIN above the outer join cannot be pushed down below the outer join.
 * Otherwise the outer join will produce wrong results because it will see the
 * wrong sets of input rows.  All quals are stored as RestrictInfo nodes
 * during planning, but there's a flag to indicate whether a qual has been
 * pushed down to a lower level than its original syntactic placement in the
 * join tree would suggest.  If an outer join prevents us from pushing a qual
 * down to its "natural" semantic level (the level associated with just the
 * base rels used in the qual) then the qual will appear in JoinInfo lists
 * that reference more than just the base rels it actually uses.  By
 * pretending that the qual references all the rels appearing in the outer
 * join, we prevent it from being evaluated below the outer join's joinrel.
 * When we do form the outer join's joinrel, we still need to distinguish
 * those quals that are actually in that join's JOIN/ON condition from those
 * that appeared higher in the tree and were pushed down to the join rel
 * because they used no other rels.  That's what the ispusheddown flag is for;
 * it tells us that a qual came from a point above the join of the specific
 * set of base rels that it uses (or that the JoinInfo structures claim it
 * uses).  A clause that originally came from WHERE will *always* have its
 * ispusheddown flag set; a clause that came from an INNER JOIN condition,
 * but doesn't use all the rels being joined, will also have ispusheddown set
 * because it will get attached to some lower joinrel.
 *
 * In general, the referenced clause might be arbitrarily complex.	The
 * kinds of clauses we can handle as indexscan quals, mergejoin clauses,
 * or hashjoin clauses are fairly limited --- the code for each kind of
 * path is responsible for identifying the restrict clauses it can use
 * and ignoring the rest.  Clauses not implemented by an indexscan,
 * mergejoin, or hashjoin will be placed in the plan qual or joinqual field
 * of the finished Plan node, where they will be enforced by general-purpose
 * qual-expression-evaluation code.  (But we are still entitled to count
 * their selectivity when estimating the result tuple count, if we
 * can guess what it is...)
 */

typedef struct RestrictInfo
{
	NodeTag		type;

	Expr	   *clause;			/* the represented clause of WHERE or JOIN */

	bool		ispusheddown;	/* TRUE if clause was pushed down in level */

	/* only used if clause is an OR clause: */
	List	   *subclauseindices;		/* indexes matching subclauses */
	/* subclauseindices is a List of Lists of IndexOptInfos */

	/* cache space for costs (currently only used for join clauses) */
	QualCost	eval_cost;		/* eval cost of clause; -1 if not yet set */
	Selectivity this_selec;		/* selectivity; -1 if not yet set */

	/*
	 * If the clause looks useful for joining --- that is, it is a binary
	 * opclause with nonoverlapping sets of relids referenced in the left
	 * and right sides --- then these two fields are set to sets of the
	 * referenced relids.  Otherwise they are both NULL.
	 */
	Relids		left_relids;	/* relids in left side of join clause */
	Relids		right_relids;	/* relids in right side of join clause */

	/* valid if clause is mergejoinable, else InvalidOid: */
	Oid			mergejoinoperator;		/* copy of clause operator */
	Oid			left_sortop;	/* leftside sortop needed for mergejoin */
	Oid			right_sortop;	/* rightside sortop needed for mergejoin */

	/* cache space for mergeclause processing; NIL if not yet set */
	List	   *left_pathkey;	/* canonical pathkey for left side */
	List	   *right_pathkey;	/* canonical pathkey for right side */

	/* cache space for mergeclause processing; -1 if not yet set */
	Selectivity left_mergescansel;		/* fraction of left side to scan */
	Selectivity right_mergescansel;		/* fraction of right side to scan */

	/* valid if clause is hashjoinable, else InvalidOid: */
	Oid			hashjoinoperator;		/* copy of clause operator */

	/* cache space for hashclause processing; -1 if not yet set */
	Selectivity left_bucketsize;	/* avg bucketsize of left side */
	Selectivity right_bucketsize;		/* avg bucketsize of right side */
} RestrictInfo;

/*
 * Join clause info.
 *
 * We make a list of these for each RelOptInfo, containing info about
 * all the join clauses this RelOptInfo participates in.  (For this
 * purpose, a "join clause" is a WHERE clause that mentions both vars
 * belonging to this relation and vars belonging to relations not yet
 * joined to it.)  We group these clauses according to the set of
 * other base relations (unjoined relations) mentioned in them.
 * There is one JoinInfo for each distinct set of unjoined_relids,
 * and its jinfo_restrictinfo lists the clause(s) that use that set
 * of other relations.
 */

typedef struct JoinInfo
{
	NodeTag		type;
	Relids		unjoined_relids;	/* some rels not yet part of my RelOptInfo */
	List	   *jinfo_restrictinfo;		/* relevant RestrictInfos */
} JoinInfo;

/*
 * Inner indexscan info.
 *
 * An inner indexscan is one that uses one or more joinclauses as index
 * conditions (perhaps in addition to plain restriction clauses).  So it
 * can only be used as the inner path of a nestloop join where the outer
 * relation includes all other relids appearing in those joinclauses.
 * The set of usable joinclauses, and thus the best inner indexscan,
 * thus varies depending on which outer relation we consider; so we have
 * to recompute the best such path for every join.	To avoid lots of
 * redundant computation, we cache the results of such searches.  For
 * each index we compute the set of possible otherrelids (all relids
 * appearing in joinquals that could become indexquals for this index).
 * Two outer relations whose relids have the same intersection with this
 * set will have the same set of available joinclauses and thus the same
 * best inner indexscan for that index.  Similarly, for each base relation,
 * we form the union of the per-index otherrelids sets.  Two outer relations
 * with the same intersection with that set will have the same best overall
 * inner indexscan for the base relation.  We use lists of InnerIndexscanInfo
 * nodes to cache the results of these searches at both the index and
 * relation level.
 *
 * The search key also includes a bool showing whether the join being
 * considered is an outer join.  Since we constrain the join order for
 * outer joins, I believe that this bool can only have one possible value
 * for any particular base relation; but store it anyway to avoid confusion.
 */

typedef struct InnerIndexscanInfo
{
	NodeTag		type;
	/* The lookup key: */
	Relids		other_relids;	/* a set of relevant other relids */
	bool		isouterjoin;	/* true if join is outer */
	/* Best path for this lookup key: */
	Path	   *best_innerpath; /* best inner indexscan, or NULL if none */
} InnerIndexscanInfo;

/*
 * IN clause info.
 *
 * When we convert top-level IN quals into join operations, we must restrict
 * the order of joining and use special join methods at some join points.
 * We record information about each such IN clause in an InClauseInfo struct.
 * These structs are kept in the Query node's in_info_list.
 */

typedef struct InClauseInfo
{
	NodeTag		type;
	Relids		lefthand;		/* base relids in lefthand expressions */
	Relids		righthand;		/* base relids coming from the subselect */
	List	   *sub_targetlist; /* targetlist of original RHS subquery */

	/*
	 * Note: sub_targetlist is just a list of Vars or expressions; it does
	 * not contain TargetEntry nodes.
	 */
} InClauseInfo;

#endif   /* RELATION_H */