Some further performance tweaks for planning large inheritance trees that

are mostly excluded by constraints: do the CE test a bit earlier to save
some adjust_appendrel_attrs() work on excluded children, and arrange to
use array indexing rather than rt_fetch() to fetch RTEs in the main body
of the planner.  The latter is something I'd wanted to do for awhile anyway,
but seeing list_nth_cell() as 35% of the runtime gets one's attention.

diff --git a/src/backend/optimizer/path/allpaths.c b/src/backend/optimizer/path/allpaths.c
index 0d87291025f5e08f7b9c447cb8db8033c74e221c..07a99e9c6b7c6503542e1346cce33ef2377aefa9 100644 (file)
--- a/src/backend/optimizer/path/allpaths.c
+++ b/src/backend/optimizer/path/allpaths.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/optimizer/path/allpaths.c,v 1.162 2007/04/21 06:18:52 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/optimizer/path/allpaths.c,v 1.163 2007/04/21 21:01:44 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -39,7 +39,8 @@ int           geqo_threshold;
 
 
 static void set_base_rel_pathlists(PlannerInfo *root);
-static void set_rel_pathlist(PlannerInfo *root, RelOptInfo *rel, Index rti);
+static void set_rel_pathlist(PlannerInfo *root, RelOptInfo *rel,
+                            Index rti, RangeTblEntry *rte);
 static void set_plain_rel_pathlist(PlannerInfo *root, RelOptInfo *rel,
                       RangeTblEntry *rte);
 static void set_append_rel_pathlist(PlannerInfo *root, RelOptInfo *rel,
@@ -144,7 +145,7 @@ set_base_rel_pathlists(PlannerInfo *root)
        if (rel->reloptkind != RELOPT_BASEREL)
            continue;
 
-       set_rel_pathlist(root, rel, rti);
+       set_rel_pathlist(root, rel, rti, root->simple_rte_array[rti]);
    }
 }
 
@@ -153,10 +154,9 @@ set_base_rel_pathlists(PlannerInfo *root)
  *   Build access paths for a base relation
  */
 static void
-set_rel_pathlist(PlannerInfo *root, RelOptInfo *rel, Index rti)
+set_rel_pathlist(PlannerInfo *root, RelOptInfo *rel,
+                Index rti, RangeTblEntry *rte)
 {
-   RangeTblEntry *rte = rt_fetch(rti, root->parse->rtable);
-
    if (rte->inh)
    {
        /* It's an "append relation", process accordingly */
@@ -200,8 +200,11 @@ set_plain_rel_pathlist(PlannerInfo *root, RelOptInfo *rel, RangeTblEntry *rte)
     * If we can prove we don't need to scan the rel via constraint exclusion,
     * set up a single dummy path for it.  (Rather than inventing a special
     * "dummy" path type, we represent this as an AppendPath with no members.)
+    * We only need to check for regular baserels; if it's an otherrel, CE
+    * was already checked in set_append_rel_pathlist().
     */
-   if (relation_excluded_by_constraints(rel, rte))
+   if (rel->reloptkind == RELOPT_BASEREL &&
+       relation_excluded_by_constraints(rel, rte))
    {
        /* Set dummy size estimates --- we leave attr_widths[] as zeroes */
        rel->rows = 0;
@@ -294,6 +297,7 @@ set_append_rel_pathlist(PlannerInfo *root, RelOptInfo *rel,
    {
        AppendRelInfo *appinfo = (AppendRelInfo *) lfirst(l);
        int         childRTindex;
+       RangeTblEntry *childRTE;
        RelOptInfo *childrel;
        Path       *childpath;
        ListCell   *parentvars;
@@ -304,6 +308,7 @@ set_append_rel_pathlist(PlannerInfo *root, RelOptInfo *rel,
            continue;
 
        childRTindex = appinfo->child_relid;
+       childRTE = root->simple_rte_array[childRTindex];
 
        /*
         * The child rel's RelOptInfo was already created during
@@ -313,18 +318,29 @@ set_append_rel_pathlist(PlannerInfo *root, RelOptInfo *rel,
        Assert(childrel->reloptkind == RELOPT_OTHER_MEMBER_REL);
 
        /*
-        * Copy the parent's targetlist and quals to the child, with
-        * appropriate substitution of variables.
+        * We have to copy the parent's targetlist and quals to the child,
+        * with appropriate substitution of variables.  However, only the
+        * baserestrictinfo quals are needed before we can check for
+        * constraint exclusion; so do that first and then check to see
+        * if we can disregard this child.
         */
-       childrel->reltargetlist = (List *)
-           adjust_appendrel_attrs((Node *) rel->reltargetlist,
-                                  appinfo);
        childrel->baserestrictinfo = (List *)
            adjust_appendrel_attrs((Node *) rel->baserestrictinfo,
                                   appinfo);
+
+       if (relation_excluded_by_constraints(childrel, childRTE))
+       {
+           /* this child need not be scanned, so just disregard it */
+           continue;
+       }
+
+       /* CE failed, so finish copying targetlist and join quals */
        childrel->joininfo = (List *)
            adjust_appendrel_attrs((Node *) rel->joininfo,
                                   appinfo);
+       childrel->reltargetlist = (List *)
+           adjust_appendrel_attrs((Node *) rel->reltargetlist,
+                                  appinfo);
 
        /*
         * We have to make child entries in the EquivalenceClass data
@@ -353,12 +369,9 @@ set_append_rel_pathlist(PlannerInfo *root, RelOptInfo *rel,
         * It's possible that the child is itself an appendrel, in which case
         * we can "cut out the middleman" and just add its child paths to our
         * own list.  (We don't try to do this earlier because we need to
-        * apply both levels of transformation to the quals.) This test also
-        * handles the case where the child rel need not be scanned because of
-        * constraint exclusion: it'll have an Append path with no subpaths,
-        * and will vanish from our list.
+        * apply both levels of transformation to the quals.)
         */
-       set_rel_pathlist(root, childrel, childRTindex);
+       set_rel_pathlist(root, childrel, childRTindex, childRTE);
 
        childpath = childrel->cheapest_total_path;
        if (IsA(childpath, AppendPath))

diff --git a/src/backend/optimizer/path/clausesel.c b/src/backend/optimizer/path/clausesel.c
index 405ab6fcc3a6275dc1a132e547693b6ce6bbde07..b8bbc29c505bb3c418144644d8d0cac819003d59 100644 (file)
--- a/src/backend/optimizer/path/clausesel.c
+++ b/src/backend/optimizer/path/clausesel.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/optimizer/path/clausesel.c,v 1.84 2007/02/19 07:03:28 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/optimizer/path/clausesel.c,v 1.85 2007/04/21 21:01:44 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -500,7 +500,7 @@ clause_selectivity(PlannerInfo *root,
        if (var->varlevelsup == 0 &&
            (varRelid == 0 || varRelid == (int) var->varno))
        {
-           RangeTblEntry *rte = rt_fetch(var->varno, root->parse->rtable);
+           RangeTblEntry *rte = planner_rt_fetch(var->varno, root);
 
            if (rte->rtekind == RTE_SUBQUERY)
            {

diff --git a/src/backend/optimizer/path/costsize.c b/src/backend/optimizer/path/costsize.c
index 42a5520b974ebf0f0947923534e349f00be16d69..9cd3a51d8976c99b9157d21341df745f1145fa4a 100644 (file)
--- a/src/backend/optimizer/path/costsize.c
+++ b/src/backend/optimizer/path/costsize.c
@@ -54,7 +54,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/optimizer/path/costsize.c,v 1.180 2007/04/21 02:41:13 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/optimizer/path/costsize.c,v 1.181 2007/04/21 21:01:44 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -856,7 +856,7 @@ cost_functionscan(Path *path, PlannerInfo *root, RelOptInfo *baserel)
 
    /* Should only be applied to base relations that are functions */
    Assert(baserel->relid > 0);
-   rte = rt_fetch(baserel->relid, root->parse->rtable);
+   rte = planner_rt_fetch(baserel->relid, root);
    Assert(rte->rtekind == RTE_FUNCTION);
 
    /* Estimate costs of executing the function expression */
@@ -2297,7 +2297,7 @@ set_function_size_estimates(PlannerInfo *root, RelOptInfo *rel)
 
    /* Should only be applied to base relations that are functions */
    Assert(rel->relid > 0);
-   rte = rt_fetch(rel->relid, root->parse->rtable);
+   rte = planner_rt_fetch(rel->relid, root);
    Assert(rte->rtekind == RTE_FUNCTION);
 
    /* Estimate number of rows the function itself will return */
@@ -2323,7 +2323,7 @@ set_values_size_estimates(PlannerInfo *root, RelOptInfo *rel)
 
    /* Should only be applied to base relations that are values lists */
    Assert(rel->relid > 0);
-   rte = rt_fetch(rel->relid, root->parse->rtable);
+   rte = planner_rt_fetch(rel->relid, root);
    Assert(rte->rtekind == RTE_VALUES);
 
    /*

diff --git a/src/backend/optimizer/plan/createplan.c b/src/backend/optimizer/plan/createplan.c
index 9b9645faf28a0d46e9b560e6dce8757f3556ff80..74b89125665f65c66094e71bfb87d7f80399ad5e 100644 (file)
--- a/src/backend/optimizer/plan/createplan.c
+++ b/src/backend/optimizer/plan/createplan.c
@@ -10,7 +10,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/optimizer/plan/createplan.c,v 1.228 2007/04/06 22:33:42 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/optimizer/plan/createplan.c,v 1.229 2007/04/21 21:01:44 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1353,7 +1353,7 @@ create_functionscan_plan(PlannerInfo *root, Path *best_path,
 
    /* it should be a function base rel... */
    Assert(scan_relid > 0);
-   rte = rt_fetch(scan_relid, root->parse->rtable);
+   rte = planner_rt_fetch(scan_relid, root);
    Assert(rte->rtekind == RTE_FUNCTION);
 
    /* Sort clauses into best execution order */
@@ -1388,7 +1388,7 @@ create_valuesscan_plan(PlannerInfo *root, Path *best_path,
 
    /* it should be a values base rel... */
    Assert(scan_relid > 0);
-   rte = rt_fetch(scan_relid, root->parse->rtable);
+   rte = planner_rt_fetch(scan_relid, root);
    Assert(rte->rtekind == RTE_VALUES);
 
    /* Sort clauses into best execution order */

diff --git a/src/backend/optimizer/plan/planagg.c b/src/backend/optimizer/plan/planagg.c
index 95f0e64cc4190e7aa680de1b3740d1eeed287398..92ade2fccf56a08831d0be2bc50e3464c98a174d 100644 (file)
--- a/src/backend/optimizer/plan/planagg.c
+++ b/src/backend/optimizer/plan/planagg.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/optimizer/plan/planagg.c,v 1.30 2007/03/17 00:11:04 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/optimizer/plan/planagg.c,v 1.31 2007/04/21 21:01:45 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -116,7 +116,7 @@ optimize_minmax_aggregates(PlannerInfo *root, List *tlist, Path *best_path)
    if (!IsA(jtnode, RangeTblRef))
        return NULL;
    rtr = (RangeTblRef *) jtnode;
-   rte = rt_fetch(rtr->rtindex, parse->rtable);
+   rte = planner_rt_fetch(rtr->rtindex, root);
    if (rte->rtekind != RTE_RELATION || rte->inh)
        return NULL;
    rel = find_base_rel(root, rtr->rtindex);

diff --git a/src/backend/optimizer/plan/planmain.c b/src/backend/optimizer/plan/planmain.c
index de3944b6b07a757f7a59e201e7d20fe12f71a620..f12e388d7235881c96a7cda4995300f8763d26f4 100644 (file)
--- a/src/backend/optimizer/plan/planmain.c
+++ b/src/backend/optimizer/plan/planmain.c
@@ -14,7 +14,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/optimizer/plan/planmain.c,v 1.99 2007/01/20 20:45:39 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/optimizer/plan/planmain.c,v 1.100 2007/04/21 21:01:45 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -86,6 +86,7 @@ query_planner(PlannerInfo *root, List *tlist, double tuple_fraction,
    Path       *cheapestpath;
    Path       *sortedpath;
    Index       rti;
+   ListCell   *lc;
    double      total_pages;
 
    /* Make tuple_fraction accessible to lower-level routines */
@@ -123,6 +124,20 @@ query_planner(PlannerInfo *root, List *tlist, double tuple_fraction,
    root->full_join_clauses = NIL;
    root->oj_info_list = NIL;
 
+   /*
+    * Make a flattened version of the rangetable for faster access (this
+    * is OK because the rangetable won't change any more).
+    */
+   root->simple_rte_array = (RangeTblEntry **)
+       palloc0(root->simple_rel_array_size * sizeof(RangeTblEntry *));
+   rti = 1;
+   foreach(lc, parse->rtable)
+   {
+       RangeTblEntry *rte = (RangeTblEntry *) lfirst(lc);
+
+       root->simple_rte_array[rti++] = rte;
+   }
+
    /*
     * Construct RelOptInfo nodes for all base relations in query, and
     * indirectly for all appendrel member relations ("other rels").  This

diff --git a/src/backend/optimizer/util/pathnode.c b/src/backend/optimizer/util/pathnode.c
index 81f7c99e96310a6584b7999bed6f5656d5b469c9..6d440001d6b2a28a0f6080cbe6805c9c91d8e66c 100644 (file)
--- a/src/backend/optimizer/util/pathnode.c
+++ b/src/backend/optimizer/util/pathnode.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/optimizer/util/pathnode.c,v 1.138 2007/02/06 02:59:12 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/optimizer/util/pathnode.c,v 1.139 2007/04/21 21:01:45 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -800,7 +800,7 @@ create_unique_path(PlannerInfo *root, RelOptInfo *rel, Path *subpath)
     */
    if (sub_targetlist && rel->rtekind == RTE_SUBQUERY)
    {
-       RangeTblEntry *rte = rt_fetch(rel->relid, root->parse->rtable);
+       RangeTblEntry *rte = planner_rt_fetch(rel->relid, root);
        List       *sub_tlist_colnos;
 
        sub_tlist_colnos = translate_sub_tlist(sub_targetlist, rel->relid);

diff --git a/src/backend/optimizer/util/plancat.c b/src/backend/optimizer/util/plancat.c
index 33b081ffffddb93b36f20def780ddaa3fb036819..662fcbaaf8ed6c448c639ea930a2fb70b5aad63d 100644 (file)
--- a/src/backend/optimizer/util/plancat.c
+++ b/src/backend/optimizer/util/plancat.c
@@ -9,7 +9,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/optimizer/util/plancat.c,v 1.133 2007/04/06 22:33:42 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/optimizer/util/plancat.c,v 1.134 2007/04/21 21:01:45 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -497,6 +497,9 @@ get_relation_constraints(Oid relationObjectId, RelOptInfo *rel)
  * Detect whether the relation need not be scanned because it has either
  * self-inconsistent restrictions, or restrictions inconsistent with the
  * relation's CHECK constraints.
+ *
+ * Note: this examines only rel->relid and rel->baserestrictinfo; therefore
+ * it can be called before filling in other fields of the RelOptInfo.
  */
 bool
 relation_excluded_by_constraints(RelOptInfo *rel, RangeTblEntry *rte)
@@ -595,7 +598,7 @@ build_physical_tlist(PlannerInfo *root, RelOptInfo *rel)
 {
    List       *tlist = NIL;
    Index       varno = rel->relid;
-   RangeTblEntry *rte = rt_fetch(varno, root->parse->rtable);
+   RangeTblEntry *rte = planner_rt_fetch(varno, root);
    Relation    relation;
    Query      *subquery;
    Var        *var;

diff --git a/src/backend/optimizer/util/relnode.c b/src/backend/optimizer/util/relnode.c
index 6dfd2f61149e71c2ef87ec965b64a1b850e04aea..56f8f3493c223e539adc91e2d4cd55164204ca4c 100644 (file)
--- a/src/backend/optimizer/util/relnode.c
+++ b/src/backend/optimizer/util/relnode.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/optimizer/util/relnode.c,v 1.86 2007/02/22 22:00:25 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/optimizer/util/relnode.c,v 1.87 2007/04/21 21:01:45 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -56,15 +56,15 @@ build_simple_rel(PlannerInfo *root, int relid, RelOptKind reloptkind)
    RelOptInfo *rel;
    RangeTblEntry *rte;
 
-   /* Fetch RTE for relation */
-   Assert(relid > 0 && relid <= list_length(root->parse->rtable));
-   rte = rt_fetch(relid, root->parse->rtable);
-
    /* Rel should not exist already */
-   Assert(relid < root->simple_rel_array_size);
+   Assert(relid > 0 && relid < root->simple_rel_array_size);
    if (root->simple_rel_array[relid] != NULL)
        elog(ERROR, "rel %d already exists", relid);
 
+   /* Fetch RTE for relation */
+   rte = root->simple_rte_array[relid];
+   Assert(rte != NULL);
+
    rel = makeNode(RelOptInfo);
    rel->reloptkind = reloptkind;
    rel->relids = bms_make_singleton(relid);

diff --git a/src/backend/utils/adt/selfuncs.c b/src/backend/utils/adt/selfuncs.c
index 28e95a74eedc0e7534fee3c695a9e5983385bd17..079f5e2cc9dfff4b12ea1c9b7e3fd5183d00bf58 100644 (file)
--- a/src/backend/utils/adt/selfuncs.c
+++ b/src/backend/utils/adt/selfuncs.c
@@ -15,7 +15,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/utils/adt/selfuncs.c,v 1.232 2007/04/06 22:33:42 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/utils/adt/selfuncs.c,v 1.233 2007/04/21 21:01:45 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -3487,7 +3487,7 @@ examine_variable(PlannerInfo *root, Node *node, int varRelid,
        vardata->atttype = var->vartype;
        vardata->atttypmod = var->vartypmod;
 
-       rte = rt_fetch(var->varno, root->parse->rtable);
+       rte = root->simple_rte_array[var->varno];
 
        if (rte->inh)
        {

diff --git a/src/include/nodes/relation.h b/src/include/nodes/relation.h
index 2c26d121ef406abeb50a16a90ffdd7e054f06cef..57af937830f0722a7d393b96fe716edc1d1443c0 100644 (file)
--- a/src/include/nodes/relation.h
+++ b/src/include/nodes/relation.h
@@ -7,7 +7,7 @@
  * Portions Copyright (c) 1996-2007, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/nodes/relation.h,v 1.140 2007/04/06 22:33:43 tgl Exp $
+ * $PostgreSQL: pgsql/src/include/nodes/relation.h,v 1.141 2007/04/21 21:01:45 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -108,6 +108,14 @@ typedef struct PlannerInfo
    struct RelOptInfo **simple_rel_array;       /* All 1-rel RelOptInfos */
    int         simple_rel_array_size;  /* allocated size of array */
 
+   /*
+    * simple_rte_array is the same length as simple_rel_array and holds
+    * pointers to the associated rangetable entries.  This lets us avoid
+    * rt_fetch(), which can be a bit slow once large inheritance sets have
+    * been expanded.
+    */
+   RangeTblEntry **simple_rte_array;           /* rangetable as an array */
+
    /*
     * join_rel_list is a list of all join-relation RelOptInfos we have
     * considered in this planning run.  For small problems we just scan the
@@ -167,6 +175,16 @@ typedef struct PlannerInfo
 } PlannerInfo;
 
 
+/*
+ * In places where it's known that simple_rte_array[] must have been prepared
+ * already, we just index into it to fetch RTEs.  In code that might be
+ * executed before or after entering query_planner(), use this macro.
+ */
+#define planner_rt_fetch(rti, root) \
+   ((root)->simple_rte_array ? (root)->simple_rte_array[rti] : \
+    rt_fetch(rti, (root)->parse->rtable))
+
+
 /*----------
  * RelOptInfo
  *     Per-relation information for planning/optimization