Restore foreign-key-aware estimation of join relation sizes.

This patch provides a new implementation of the logic added by commit
137805f89 and later removed by 77ba61080.  It differs from the original
primarily in expending much less effort per joinrel in large queries,
which it accomplishes by doing most of the matching work once per query not
once per joinrel.  Hopefully, it's also less buggy and better commented.
The never-documented enable_fkey_estimates GUC remains gone.

There remains work to be done to make the selectivity estimates account
for nulls in FK referencing columns; but that was true of the original
patch as well.  We may be able to address this point later in beta.
In the meantime, any error should be in the direction of overestimating
rather than underestimating joinrel sizes, which seems like the direction
we want to err in.

Tomas Vondra and Tom Lane

Discussion: <31041.1465069446@sss.pgh.pa.us>

src/include/nodes/nodes.h

@@ -223,6 +223,7 @@ typedef enum NodeTag
 	T_PlannerGlobal,
 	T_RelOptInfo,
 	T_IndexOptInfo,
+	T_ForeignKeyOptInfo,
 	T_ParamPathInfo,
 	T_Path,
 	T_IndexPath,
@@ -478,7 +479,8 @@ typedef enum NodeTag
 	T_InlineCodeBlock,			/* in nodes/parsenodes.h */
 	T_FdwRoutine,				/* in foreign/fdwapi.h */
 	T_IndexAmRoutine,			/* in access/amapi.h */
-	T_TsmRoutine				/* in access/tsmapi.h */
+	T_TsmRoutine,				/* in access/tsmapi.h */
+	T_ForeignKeyCacheInfo		/* in utils/rel.h */
 } NodeTag;
 
 /*

src/include/nodes/relation.h

@@ -251,6 +251,8 @@ typedef struct PlannerInfo
 
 	List	   *placeholder_list;		/* list of PlaceHolderInfos */
 
+	List	   *fkey_list;		/* list of ForeignKeyOptInfos */
+
 	List	   *query_pathkeys; /* desired pathkeys for query_planner() */
 
 	List	   *group_pathkeys; /* groupClause pathkeys, if any */
@@ -622,6 +624,36 @@ typedef struct IndexOptInfo
 	void		(*amcostestimate) ();	/* AM's cost estimator */
 } IndexOptInfo;
 
+/*
+ * ForeignKeyOptInfo
+ *		Per-foreign-key information for planning/optimization
+ *
+ * The per-FK-column arrays can be fixed-size because we allow at most
+ * INDEX_MAX_KEYS columns in a foreign key constraint.  Each array has
+ * nkeys valid entries.
+ */
+typedef struct ForeignKeyOptInfo
+{
+	NodeTag		type;
+
+	/* Basic data about the foreign key (fetched from catalogs): */
+	Index		con_relid;		/* RT index of the referencing table */
+	Index		ref_relid;		/* RT index of the referenced table */
+	int			nkeys;			/* number of columns in the foreign key */
+	AttrNumber	conkey[INDEX_MAX_KEYS]; /* cols in referencing table */
+	AttrNumber	confkey[INDEX_MAX_KEYS];		/* cols in referenced table */
+	Oid			conpfeqop[INDEX_MAX_KEYS];		/* PK = FK operator OIDs */
+
+	/* Derived info about whether FK's equality conditions match the query: */
+	int			nmatched_ec;	/* # of FK cols matched by ECs */
+	int			nmatched_rcols; /* # of FK cols matched by non-EC rinfos */
+	int			nmatched_ri;	/* total # of non-EC rinfos matched to FK */
+	/* Pointer to eclass matching each column's condition, if there is one */
+	struct EquivalenceClass *eclass[INDEX_MAX_KEYS];
+	/* List of non-EC RestrictInfos matching each column's condition */
+	List	   *rinfos[INDEX_MAX_KEYS];
+} ForeignKeyOptInfo;
+
 
 /*
  * EquivalenceClasses

src/include/optimizer/cost.h

@@ -167,8 +167,8 @@ extern double get_parameterized_baserel_size(PlannerInfo *root,
 							   List *param_clauses);
 extern double get_parameterized_joinrel_size(PlannerInfo *root,
 							   RelOptInfo *rel,
-							   double outer_rows,
-							   double inner_rows,
+							   Path *outer_path,
+							   Path *inner_path,
 							   SpecialJoinInfo *sjinfo,
 							   List *restrict_clauses);
 extern void set_joinrel_size_estimates(PlannerInfo *root, RelOptInfo *rel,

src/include/optimizer/paths.h

@@ -140,6 +140,9 @@ extern List *generate_join_implied_equalities_for_ecs(PlannerInfo *root,
 										 Relids outer_relids,
 										 RelOptInfo *inner_rel);
 extern bool exprs_known_equal(PlannerInfo *root, Node *item1, Node *item2);
+extern EquivalenceClass *match_eclasses_to_foreign_key_col(PlannerInfo *root,
+								  ForeignKeyOptInfo *fkinfo,
+								  int colno);
 extern void add_child_rel_equivalences(PlannerInfo *root,
 						   AppendRelInfo *appinfo,
 						   RelOptInfo *parent_rel,

src/include/optimizer/planmain.h

@@ -95,6 +95,7 @@ extern RestrictInfo *build_implied_join_equality(Oid opno,
 							Expr *item2,
 							Relids qualscope,
 							Relids nullable_relids);
+extern void match_foreign_keys_to_quals(PlannerInfo *root);
 
 /*
  * prototypes for plan/analyzejoins.c

src/include/utils/rel.h

@@ -90,6 +90,10 @@ typedef struct RelationData
 	/* use "struct" here to avoid needing to include rowsecurity.h: */
 	struct RowSecurityDesc *rd_rsdesc;	/* row security policies, or NULL */
 
+	/* data managed by RelationGetFKeyList: */
+	List	   *rd_fkeylist;	/* list of ForeignKeyCacheInfo (see below) */
+	bool		rd_fkeyvalid;	/* true if list has been computed */
+
 	/* data managed by RelationGetIndexList: */
 	List	   *rd_indexlist;	/* list of OIDs of indexes on relation */
 	Oid			rd_oidindex;	/* OID of unique index on OID, if any */
@@ -170,6 +174,34 @@ typedef struct RelationData
 	struct PgStat_TableStatus *pgstat_info;		/* statistics collection area */
 } RelationData;
 
+
+/*
+ * ForeignKeyCacheInfo
+ *		Information the relcache can cache about foreign key constraints
+ *
+ * This is basically just an image of relevant columns from pg_constraint.
+ * We make it a subclass of Node so that copyObject() can be used on a list
+ * of these, but we also ensure it is a "flat" object without substructure,
+ * so that list_free_deep() is sufficient to free such a list.
+ * The per-FK-column arrays can be fixed-size because we allow at most
+ * INDEX_MAX_KEYS columns in a foreign key constraint.
+ *
+ * Currently, we only cache fields of interest to the planner, but the
+ * set of fields could be expanded in future.
+ */
+typedef struct ForeignKeyCacheInfo
+{
+	NodeTag		type;
+	Oid			conrelid;		/* relation constrained by the foreign key */
+	Oid			confrelid;		/* relation referenced by the foreign key */
+	int			nkeys;			/* number of columns in the foreign key */
+	/* these arrays each have nkeys valid entries: */
+	AttrNumber	conkey[INDEX_MAX_KEYS]; /* cols in referencing table */
+	AttrNumber	confkey[INDEX_MAX_KEYS];		/* cols in referenced table */
+	Oid			conpfeqop[INDEX_MAX_KEYS];		/* PK = FK operator OIDs */
+} ForeignKeyCacheInfo;
+
+
 /*
  * StdRdOptions
  *		Standard contents of rd_options for heaps and generic indexes.

src/include/utils/relcache.h

@@ -37,6 +37,7 @@ extern void RelationClose(Relation relation);
 /*
  * Routines to compute/retrieve additional cached information
  */
+extern List *RelationGetFKeyList(Relation relation);
 extern List *RelationGetIndexList(Relation relation);
 extern Oid	RelationGetOidIndex(Relation relation);
 extern Oid	RelationGetReplicaIndex(Relation relation);