X-Git-Url: http://git.osdn.net/view?a=blobdiff_plain;f=core.c;h=965b393d400103597ca0d353a7a6fec32e33bbc7;hb=8241ef0732b65ff8635eb5cc56860b2e9db40ab7;hp=b06e54c2111d29f2d8ce0d8d02a6ea5e2c6f575a;hpb=f1d323a5ae3d3175f0f4e16f89a68086f047c902;p=pghintplan%2Fpg_hint_plan.git diff --git a/core.c b/core.c index b06e54c..965b393 100644 --- a/core.c +++ b/core.c @@ -3,28 +3,85 @@ * core.c * Routines copied from PostgreSQL core distribution. * + * The main purpose of this files is having access to static functions in core. + * Another purpose is tweaking functions behavior by replacing part of them by + * macro definitions. See at the end of pg_hint_plan.c for details. Anyway, + * this file *must* contain required functions without making any change. + * + * This file contains the following functions from corresponding files. + * * src/backend/optimizer/path/allpaths.c - * set_append_rel_pathlist() - * generate_mergeappend_paths() - * accumulate_append_subpath() - * standard_join_search() + * + * static functions: + * set_plain_rel_pathlist() + * add_paths_to_append_rel() + * try_partitionwise_join() + * + * public functions: + * standard_join_search(): This funcion is not static. The reason for + * including this function is make_rels_by_clause_joins. In order to + * avoid generating apparently unwanted join combination, we decided to + * change the behavior of make_join_rel, which is called under this + * function. * * src/backend/optimizer/path/joinrels.c - * join_search_one_level() + * + * public functions: + * join_search_one_level(): We have to modify this to call my definition of + * make_rels_by_clause_joins. + * + * static functions: * make_rels_by_clause_joins() * make_rels_by_clauseless_joins() * join_is_legal() * has_join_restriction() * is_dummy_rel() - * mark_dummy_rel() * restriction_is_constant_false() + * update_child_rel_info() + * build_child_join_sjinfo() * - * Portions Copyright (c) 1996-2012, PostgreSQL Global Development Group + * + * Portions Copyright (c) 1996-2019, PostgreSQL Global Development Group * Portions Copyright (c) 1994, Regents of the University of California * *------------------------------------------------------------------------- */ +static void populate_joinrel_with_paths(PlannerInfo *root, RelOptInfo *rel1, + RelOptInfo *rel2, RelOptInfo *joinrel, + SpecialJoinInfo *sjinfo, List *restrictlist); + +/* + * set_plain_rel_pathlist + * Build access paths for a plain relation (no subquery, no inheritance) + */ +static void +set_plain_rel_pathlist(PlannerInfo *root, RelOptInfo *rel, RangeTblEntry *rte) +{ + Relids required_outer; + + /* + * We don't support pushing join clauses into the quals of a seqscan, but + * it could still have required parameterization due to LATERAL refs in + * its tlist. + */ + required_outer = rel->lateral_relids; + + /* Consider sequential scan */ + add_path(rel, create_seqscan_path(root, rel, required_outer, 0)); + + /* If appropriate, consider parallel sequential scan */ + if (rel->consider_parallel && required_outer == NULL) + create_plain_partial_paths(root, rel); + + /* Consider index scans */ + create_index_paths(root, rel); + + /* Consider TID scans */ + create_tidscan_paths(root, rel); +} + + /* * set_append_rel_pathlist * Build access paths for an "append relation" @@ -35,16 +92,11 @@ set_append_rel_pathlist(PlannerInfo *root, RelOptInfo *rel, { int parentRTindex = rti; List *live_childrels = NIL; - List *subpaths = NIL; - List *all_child_pathkeys = NIL; - List *all_child_outers = NIL; ListCell *l; /* * Generate access paths for each member relation, and remember the - * cheapest path for each one. Also, identify all pathkeys (orderings) - * and parameterizations (required_outer sets) available for the member - * relations. + * non-dummy children. */ foreach(l, root->append_rel_list) { @@ -52,7 +104,6 @@ set_append_rel_pathlist(PlannerInfo *root, RelOptInfo *rel, int childRTindex; RangeTblEntry *childRTE; RelOptInfo *childrel; - ListCell *lcp; /* append_rel_list contains all append rels; ignore others */ if (appinfo->parent_relid != parentRTindex) @@ -64,6 +115,15 @@ set_append_rel_pathlist(PlannerInfo *root, RelOptInfo *rel, childrel = root->simple_rel_array[childRTindex]; /* + * If set_append_rel_size() decided the parent appendrel was + * parallel-unsafe at some point after visiting this child rel, we + * need to propagate the unsafety marking down to the child, so that + * we don't generate useless partial paths for it. + */ + if (!rel->consider_parallel) + childrel->consider_parallel = false; + + /* * Compute the child's access paths. */ set_rel_pathlist(root, childrel, childRTindex, childRTE); @@ -74,270 +134,22 @@ set_append_rel_pathlist(PlannerInfo *root, RelOptInfo *rel, if (IS_DUMMY_REL(childrel)) continue; - /* - * Child is live, so add its cheapest access path to the Append path - * we are constructing for the parent. - */ - subpaths = accumulate_append_subpath(subpaths, - childrel->cheapest_total_path); - - /* Remember which childrels are live, for logic below */ - live_childrels = lappend(live_childrels, childrel); + /* Bubble up childrel's partitioned children. */ + if (rel->part_scheme) + rel->partitioned_child_rels = + list_concat(rel->partitioned_child_rels, + list_copy(childrel->partitioned_child_rels)); /* - * Collect lists of all the available path orderings and - * parameterizations for all the children. We use these as a - * heuristic to indicate which sort orderings and parameterizations we - * should build Append and MergeAppend paths for. + * Child is live, so add it to the live_childrels list for use below. */ - foreach(lcp, childrel->pathlist) - { - Path *childpath = (Path *) lfirst(lcp); - List *childkeys = childpath->pathkeys; - Relids childouter = PATH_REQ_OUTER(childpath); - - /* Unsorted paths don't contribute to pathkey list */ - if (childkeys != NIL) - { - ListCell *lpk; - bool found = false; - - /* Have we already seen this ordering? */ - foreach(lpk, all_child_pathkeys) - { - List *existing_pathkeys = (List *) lfirst(lpk); - - if (compare_pathkeys(existing_pathkeys, - childkeys) == PATHKEYS_EQUAL) - { - found = true; - break; - } - } - if (!found) - { - /* No, so add it to all_child_pathkeys */ - all_child_pathkeys = lappend(all_child_pathkeys, - childkeys); - } - } - - /* Unparameterized paths don't contribute to param-set list */ - if (childouter) - { - ListCell *lco; - bool found = false; - - /* Have we already seen this param set? */ - foreach(lco, all_child_outers) - { - Relids existing_outers = (Relids) lfirst(lco); - - if (bms_equal(existing_outers, childouter)) - { - found = true; - break; - } - } - if (!found) - { - /* No, so add it to all_child_outers */ - all_child_outers = lappend(all_child_outers, - childouter); - } - } - } - } - - /* - * Next, build an unordered, unparameterized Append path for the rel. - * (Note: this is correct even if we have zero or one live subpath due to - * constraint exclusion.) - */ - add_path(rel, (Path *) create_append_path(rel, subpaths, NULL)); - - /* - * Build unparameterized MergeAppend paths based on the collected list of - * child pathkeys. - */ - generate_mergeappend_paths(root, rel, live_childrels, all_child_pathkeys); - - /* - * Build Append paths for each parameterization seen among the child rels. - * (This may look pretty expensive, but in most cases of practical - * interest, the child rels will expose mostly the same parameterizations, - * so that not that many cases actually get considered here.) - * - * The Append node itself cannot enforce quals, so all qual checking must - * be done in the child paths. This means that to have a parameterized - * Append path, we must have the exact same parameterization for each - * child path; otherwise some children might be failing to check the - * moved-down quals. To make them match up, we can try to increase the - * parameterization of lesser-parameterized paths. - */ - foreach(l, all_child_outers) - { - Relids required_outer = (Relids) lfirst(l); - bool ok = true; - ListCell *lcr; - - /* Select the child paths for an Append with this parameterization */ - subpaths = NIL; - foreach(lcr, live_childrels) - { - RelOptInfo *childrel = (RelOptInfo *) lfirst(lcr); - Path *cheapest_total; - - cheapest_total = - get_cheapest_path_for_pathkeys(childrel->pathlist, - NIL, - required_outer, - TOTAL_COST); - Assert(cheapest_total != NULL); - - /* Children must have exactly the desired parameterization */ - if (!bms_equal(PATH_REQ_OUTER(cheapest_total), required_outer)) - { - cheapest_total = reparameterize_path(root, cheapest_total, - required_outer, 1.0); - if (cheapest_total == NULL) - { - ok = false; - break; - } - } - - subpaths = accumulate_append_subpath(subpaths, cheapest_total); - } - - if (ok) - add_path(rel, (Path *) - create_append_path(rel, subpaths, required_outer)); + live_childrels = lappend(live_childrels, childrel); } - /* Select cheapest paths */ - set_cheapest(rel); -} - -/* - * generate_mergeappend_paths - * Generate MergeAppend paths for an append relation - * - * Generate a path for each ordering (pathkey list) appearing in - * all_child_pathkeys. - * - * We consider both cheapest-startup and cheapest-total cases, ie, for each - * interesting ordering, collect all the cheapest startup subpaths and all the - * cheapest total paths, and build a MergeAppend path for each case. - * - * We don't currently generate any parameterized MergeAppend paths. While - * it would not take much more code here to do so, it's very unclear that it - * is worth the planning cycles to investigate such paths: there's little - * use for an ordered path on the inside of a nestloop. In fact, it's likely - * that the current coding of add_path would reject such paths out of hand, - * because add_path gives no credit for sort ordering of parameterized paths, - * and a parameterized MergeAppend is going to be more expensive than the - * corresponding parameterized Append path. If we ever try harder to support - * parameterized mergejoin plans, it might be worth adding support for - * parameterized MergeAppends to feed such joins. (See notes in - * optimizer/README for why that might not ever happen, though.) - */ -static void -generate_mergeappend_paths(PlannerInfo *root, RelOptInfo *rel, - List *live_childrels, - List *all_child_pathkeys) -{ - ListCell *lcp; - - foreach(lcp, all_child_pathkeys) - { - List *pathkeys = (List *) lfirst(lcp); - List *startup_subpaths = NIL; - List *total_subpaths = NIL; - bool startup_neq_total = false; - ListCell *lcr; - - /* Select the child paths for this ordering... */ - foreach(lcr, live_childrels) - { - RelOptInfo *childrel = (RelOptInfo *) lfirst(lcr); - Path *cheapest_startup, - *cheapest_total; - - /* Locate the right paths, if they are available. */ - cheapest_startup = - get_cheapest_path_for_pathkeys(childrel->pathlist, - pathkeys, - NULL, - STARTUP_COST); - cheapest_total = - get_cheapest_path_for_pathkeys(childrel->pathlist, - pathkeys, - NULL, - TOTAL_COST); - - /* - * If we can't find any paths with the right order just use the - * cheapest-total path; we'll have to sort it later. - */ - if (cheapest_startup == NULL || cheapest_total == NULL) - { - cheapest_startup = cheapest_total = - childrel->cheapest_total_path; - Assert(cheapest_total != NULL); - } - - /* - * Notice whether we actually have different paths for the - * "cheapest" and "total" cases; frequently there will be no point - * in two create_merge_append_path() calls. - */ - if (cheapest_startup != cheapest_total) - startup_neq_total = true; - - startup_subpaths = - accumulate_append_subpath(startup_subpaths, cheapest_startup); - total_subpaths = - accumulate_append_subpath(total_subpaths, cheapest_total); - } - - /* ... and build the MergeAppend paths */ - add_path(rel, (Path *) create_merge_append_path(root, - rel, - startup_subpaths, - pathkeys, - NULL)); - if (startup_neq_total) - add_path(rel, (Path *) create_merge_append_path(root, - rel, - total_subpaths, - pathkeys, - NULL)); - } + /* Add paths to the append relation. */ + add_paths_to_append_rel(root, rel, live_childrels); } -/* - * accumulate_append_subpath - * Add a subpath to the list being built for an Append or MergeAppend - * - * It's possible that the child is itself an Append path, 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.) - */ -static List * -accumulate_append_subpath(List *subpaths, Path *path) -{ - if (IsA(path, AppendPath)) - { - AppendPath *apath = (AppendPath *) path; - - /* list_copy is important here to avoid sharing list substructure */ - return list_concat(subpaths, list_copy(apath->subpaths)); - } - else - return lappend(subpaths, path); -} /* * standard_join_search @@ -348,7 +160,7 @@ accumulate_append_subpath(List *subpaths, Path *path) * independent jointree items in the query. This is > 1. * * 'initial_rels' is a list of RelOptInfo nodes for each independent - * jointree item. These are the components to be joined together. + * jointree item. These are the components to be joined together. * Note that levels_needed == list_length(initial_rels). * * Returns the final level of join relations, i.e., the relation that is @@ -364,7 +176,7 @@ accumulate_append_subpath(List *subpaths, Path *path) * needed for these paths need have been instantiated. * * Note to plugin authors: the functions invoked during standard_join_search() - * modify root->join_rel_list and root->join_rel_hash. If you want to do more + * modify root->join_rel_list and root->join_rel_hash. If you want to do more * than one join-order search, you'll probably need to save and restore the * original states of those data structures. See geqo_eval() for an example. */ @@ -407,12 +219,29 @@ standard_join_search(PlannerInfo *root, int levels_needed, List *initial_rels) join_search_one_level(root, lev); /* - * Do cleanup work on each just-processed rel. + * Run generate_partitionwise_join_paths() and generate_gather_paths() + * for each just-processed joinrel. We could not do this earlier + * because both regular and partial paths can get added to a + * particular joinrel at multiple times within join_search_one_level. + * + * After that, we're done creating paths for the joinrel, so run + * set_cheapest(). */ foreach(lc, root->join_rel_level[lev]) { rel = (RelOptInfo *) lfirst(lc); + /* Create paths for partitionwise joins. */ + generate_partitionwise_join_paths(root, rel); + + /* + * Except for the topmost scan/join rel, consider gathering + * partial paths. We'll do the same for the topmost scan/join rel + * once we know the final targetlist (see grouping_planner). + */ + if (lev < levels_needed) + generate_gather_paths(root, rel, false); + /* Find and save the cheapest paths for this rel */ set_cheapest(rel); @@ -437,6 +266,27 @@ standard_join_search(PlannerInfo *root, int levels_needed, List *initial_rels) } /* + * create_plain_partial_paths + * Build partial access paths for parallel scan of a plain relation + */ +static void +create_plain_partial_paths(PlannerInfo *root, RelOptInfo *rel) +{ + int parallel_workers; + + parallel_workers = compute_parallel_worker(rel, rel->pages, -1, + max_parallel_workers_per_gather); + + /* If any limit was set to zero, the user doesn't want a parallel scan. */ + if (parallel_workers <= 0) + return; + + /* Add an unordered partial path based on a parallel sequential scan. */ + add_partial_path(rel, create_seqscan_path(root, rel, NULL, parallel_workers)); +} + + +/* * join_search_one_level * Consider ways to produce join relations containing exactly 'level' * jointree items. (This is one step of the dynamic-programming method @@ -492,7 +342,7 @@ join_search_one_level(PlannerInfo *root, int level) if (level == 2) /* consider remaining initial rels */ other_rels = lnext(r); - else /* consider all initial rels */ + else /* consider all initial rels */ other_rels = list_head(joinrels[1]); make_rels_by_clause_joins(root, @@ -615,7 +465,7 @@ join_search_one_level(PlannerInfo *root, int level) /*---------- * When special joins are involved, there may be no legal way - * to make an N-way join for some values of N. For example consider + * to make an N-way join for some values of N. For example consider * * SELECT ... FROM t1 WHERE * x IN (SELECT ... FROM t2,t3 WHERE ...) AND @@ -626,15 +476,19 @@ join_search_one_level(PlannerInfo *root, int level) * to accept failure at level 4 and go on to discover a workable * bushy plan at level 5. * - * However, if there are no special joins then join_is_legal() should - * never fail, and so the following sanity check is useful. + * However, if there are no special joins and no lateral references + * then join_is_legal() should never fail, and so the following sanity + * check is useful. *---------- */ - if (joinrels[level] == NIL && root->join_info_list == NIL) + if (joinrels[level] == NIL && + root->join_info_list == NIL && + !root->hasLateralRTEs) elog(ERROR, "failed to build any %d-way joins", level); } } + /* * make_rels_by_clause_joins * Build joins between the given relation 'old_rel' and other relations @@ -675,6 +529,7 @@ make_rels_by_clause_joins(PlannerInfo *root, } } + /* * make_rels_by_clauseless_joins * Given a relation 'old_rel' and a list of other relations @@ -707,6 +562,7 @@ make_rels_by_clauseless_joins(PlannerInfo *root, } } + /* * join_is_legal * Determine whether a proposed join is legal given the query's @@ -718,7 +574,7 @@ make_rels_by_clauseless_joins(PlannerInfo *root, * * On success, *sjinfo_p is set to NULL if this is to be a plain inner join, * else it's set to point to the associated SpecialJoinInfo node. Also, - * *reversed_p is set TRUE if the given relations need to be swapped to + * *reversed_p is set true if the given relations need to be swapped to * match the SpecialJoinInfo node. */ static bool @@ -729,11 +585,11 @@ join_is_legal(PlannerInfo *root, RelOptInfo *rel1, RelOptInfo *rel2, SpecialJoinInfo *match_sjinfo; bool reversed; bool unique_ified; - bool is_valid_inner; + bool must_be_leftjoin; ListCell *l; /* - * Ensure output params are set on failure return. This is just to + * Ensure output params are set on failure return. This is just to * suppress uninitialized-variable warnings from overly anal compilers. */ *sjinfo_p = NULL; @@ -741,13 +597,13 @@ join_is_legal(PlannerInfo *root, RelOptInfo *rel1, RelOptInfo *rel2, /* * If we have any special joins, the proposed join might be illegal; and - * in any case we have to determine its join type. Scan the join info - * list for conflicts. + * in any case we have to determine its join type. Scan the join info + * list for matches and conflicts. */ match_sjinfo = NULL; reversed = false; unique_ified = false; - is_valid_inner = true; + must_be_leftjoin = false; foreach(l, root->join_info_list) { @@ -798,7 +654,8 @@ join_is_legal(PlannerInfo *root, RelOptInfo *rel1, RelOptInfo *rel2, * If one input contains min_lefthand and the other contains * min_righthand, then we can perform the SJ at this join. * - * Barf if we get matches to more than one SJ (is that possible?) + * Reject if we get matches to more than one SJ; that implies we're + * considering something that's not really valid. */ if (bms_is_subset(sjinfo->min_lefthand, rel1->relids) && bms_is_subset(sjinfo->min_righthand, rel2->relids)) @@ -863,66 +720,187 @@ join_is_legal(PlannerInfo *root, RelOptInfo *rel1, RelOptInfo *rel2, } else { - /*---------- - * Otherwise, the proposed join overlaps the RHS but isn't - * a valid implementation of this SJ. It might still be - * a legal join, however. If both inputs overlap the RHS, - * assume that it's OK. Since the inputs presumably got past - * this function's checks previously, they can't overlap the - * LHS and their violations of the RHS boundary must represent - * SJs that have been determined to commute with this one. - * We have to allow this to work correctly in cases like - * (a LEFT JOIN (b JOIN (c LEFT JOIN d))) - * when the c/d join has been determined to commute with the join - * to a, and hence d is not part of min_righthand for the upper - * join. It should be legal to join b to c/d but this will appear - * as a violation of the upper join's RHS. - * Furthermore, if one input overlaps the RHS and the other does - * not, we should still allow the join if it is a valid - * implementation of some other SJ. We have to allow this to - * support the associative identity - * (a LJ b on Pab) LJ c ON Pbc = a LJ (b LJ c ON Pbc) on Pab - * since joining B directly to C violates the lower SJ's RHS. - * We assume that make_outerjoininfo() set things up correctly - * so that we'll only match to some SJ if the join is valid. - * Set flag here to check at bottom of loop. - *---------- + /* + * Otherwise, the proposed join overlaps the RHS but isn't a valid + * implementation of this SJ. But don't panic quite yet: the RHS + * violation might have occurred previously, in one or both input + * relations, in which case we must have previously decided that + * it was OK to commute some other SJ with this one. If we need + * to perform this join to finish building up the RHS, rejecting + * it could lead to not finding any plan at all. (This can occur + * because of the heuristics elsewhere in this file that postpone + * clauseless joins: we might not consider doing a clauseless join + * within the RHS until after we've performed other, validly + * commutable SJs with one or both sides of the clauseless join.) + * This consideration boils down to the rule that if both inputs + * overlap the RHS, we can allow the join --- they are either + * fully within the RHS, or represent previously-allowed joins to + * rels outside it. */ - if (sjinfo->jointype != JOIN_SEMI && - bms_overlap(rel1->relids, sjinfo->min_righthand) && + if (bms_overlap(rel1->relids, sjinfo->min_righthand) && bms_overlap(rel2->relids, sjinfo->min_righthand)) - { - /* seems OK */ - Assert(!bms_overlap(joinrelids, sjinfo->min_lefthand)); - } - else - is_valid_inner = false; + continue; /* assume valid previous violation of RHS */ + + /* + * The proposed join could still be legal, but only if we're + * allowed to associate it into the RHS of this SJ. That means + * this SJ must be a LEFT join (not SEMI or ANTI, and certainly + * not FULL) and the proposed join must not overlap the LHS. + */ + if (sjinfo->jointype != JOIN_LEFT || + bms_overlap(joinrelids, sjinfo->min_lefthand)) + return false; /* invalid join path */ + + /* + * To be valid, the proposed join must be a LEFT join; otherwise + * it can't associate into this SJ's RHS. But we may not yet have + * found the SpecialJoinInfo matching the proposed join, so we + * can't test that yet. Remember the requirement for later. + */ + must_be_leftjoin = true; } } /* - * Fail if violated some SJ's RHS and didn't match to another SJ. However, - * "matching" to a semijoin we are implementing by unique-ification - * doesn't count (think: it's really an inner join). + * Fail if violated any SJ's RHS and didn't match to a LEFT SJ: the + * proposed join can't associate into an SJ's RHS. + * + * Also, fail if the proposed join's predicate isn't strict; we're + * essentially checking to see if we can apply outer-join identity 3, and + * that's a requirement. (This check may be redundant with checks in + * make_outerjoininfo, but I'm not quite sure, and it's cheap to test.) */ - if (!is_valid_inner && - (match_sjinfo == NULL || unique_ified)) + if (must_be_leftjoin && + (match_sjinfo == NULL || + match_sjinfo->jointype != JOIN_LEFT || + !match_sjinfo->lhs_strict)) return false; /* invalid join path */ + /* + * We also have to check for constraints imposed by LATERAL references. + */ + if (root->hasLateralRTEs) + { + bool lateral_fwd; + bool lateral_rev; + Relids join_lateral_rels; + + /* + * The proposed rels could each contain lateral references to the + * other, in which case the join is impossible. If there are lateral + * references in just one direction, then the join has to be done with + * a nestloop with the lateral referencer on the inside. If the join + * matches an SJ that cannot be implemented by such a nestloop, the + * join is impossible. + * + * Also, if the lateral reference is only indirect, we should reject + * the join; whatever rel(s) the reference chain goes through must be + * joined to first. + * + * Another case that might keep us from building a valid plan is the + * implementation restriction described by have_dangerous_phv(). + */ + lateral_fwd = bms_overlap(rel1->relids, rel2->lateral_relids); + lateral_rev = bms_overlap(rel2->relids, rel1->lateral_relids); + if (lateral_fwd && lateral_rev) + return false; /* have lateral refs in both directions */ + if (lateral_fwd) + { + /* has to be implemented as nestloop with rel1 on left */ + if (match_sjinfo && + (reversed || + unique_ified || + match_sjinfo->jointype == JOIN_FULL)) + return false; /* not implementable as nestloop */ + /* check there is a direct reference from rel2 to rel1 */ + if (!bms_overlap(rel1->relids, rel2->direct_lateral_relids)) + return false; /* only indirect refs, so reject */ + /* check we won't have a dangerous PHV */ + if (have_dangerous_phv(root, rel1->relids, rel2->lateral_relids)) + return false; /* might be unable to handle required PHV */ + } + else if (lateral_rev) + { + /* has to be implemented as nestloop with rel2 on left */ + if (match_sjinfo && + (!reversed || + unique_ified || + match_sjinfo->jointype == JOIN_FULL)) + return false; /* not implementable as nestloop */ + /* check there is a direct reference from rel1 to rel2 */ + if (!bms_overlap(rel2->relids, rel1->direct_lateral_relids)) + return false; /* only indirect refs, so reject */ + /* check we won't have a dangerous PHV */ + if (have_dangerous_phv(root, rel2->relids, rel1->lateral_relids)) + return false; /* might be unable to handle required PHV */ + } + + /* + * LATERAL references could also cause problems later on if we accept + * this join: if the join's minimum parameterization includes any rels + * that would have to be on the inside of an outer join with this join + * rel, then it's never going to be possible to build the complete + * query using this join. We should reject this join not only because + * it'll save work, but because if we don't, the clauseless-join + * heuristics might think that legality of this join means that some + * other join rel need not be formed, and that could lead to failure + * to find any plan at all. We have to consider not only rels that + * are directly on the inner side of an OJ with the joinrel, but also + * ones that are indirectly so, so search to find all such rels. + */ + join_lateral_rels = min_join_parameterization(root, joinrelids, + rel1, rel2); + if (join_lateral_rels) + { + Relids join_plus_rhs = bms_copy(joinrelids); + bool more; + + do + { + more = false; + foreach(l, root->join_info_list) + { + SpecialJoinInfo *sjinfo = (SpecialJoinInfo *) lfirst(l); + + if (bms_overlap(sjinfo->min_lefthand, join_plus_rhs) && + !bms_is_subset(sjinfo->min_righthand, join_plus_rhs)) + { + join_plus_rhs = bms_add_members(join_plus_rhs, + sjinfo->min_righthand); + more = true; + } + /* full joins constrain both sides symmetrically */ + if (sjinfo->jointype == JOIN_FULL && + bms_overlap(sjinfo->min_righthand, join_plus_rhs) && + !bms_is_subset(sjinfo->min_lefthand, join_plus_rhs)) + { + join_plus_rhs = bms_add_members(join_plus_rhs, + sjinfo->min_lefthand); + more = true; + } + } + } while (more); + if (bms_overlap(join_plus_rhs, join_lateral_rels)) + return false; /* will not be able to join to some RHS rel */ + } + } + /* Otherwise, it's a valid join */ *sjinfo_p = match_sjinfo; *reversed_p = reversed; return true; } + /* * has_join_restriction - * Detect whether the specified relation has join-order restrictions - * due to being inside an outer join or an IN (sub-SELECT). + * Detect whether the specified relation has join-order restrictions, + * due to being inside an outer join or an IN (sub-SELECT), + * or participating in any LATERAL references or multi-rel PHVs. * * Essentially, this tests whether have_join_order_restriction() could * succeed with this rel and some other one. It's OK if we sometimes - * say "true" incorrectly. (Therefore, we don't bother with the relatively + * say "true" incorrectly. (Therefore, we don't bother with the relatively * expensive has_legal_joinclause test.) */ static bool @@ -930,6 +908,18 @@ has_join_restriction(PlannerInfo *root, RelOptInfo *rel) { ListCell *l; + if (rel->lateral_relids != NULL || rel->lateral_referencers != NULL) + return true; + + foreach(l, root->placeholder_list) + { + PlaceHolderInfo *phinfo = (PlaceHolderInfo *) lfirst(l); + + if (bms_is_subset(rel->relids, phinfo->ph_eval_at) && + !bms_equal(rel->relids, phinfo->ph_eval_at)) + return true; + } + foreach(l, root->join_info_list) { SpecialJoinInfo *sjinfo = (SpecialJoinInfo *) lfirst(l); @@ -952,6 +942,7 @@ has_join_restriction(PlannerInfo *root, RelOptInfo *rel) return false; } + /* * is_dummy_rel --- has relation been proven empty? */ @@ -969,14 +960,14 @@ is_dummy_rel(RelOptInfo *rel) * dummy. * * Also, when called during GEQO join planning, we are in a short-lived - * memory context. We must make sure that the dummy path attached to a + * memory context. We must make sure that the dummy path attached to a * baserel survives the GEQO cycle, else the baserel is trashed for future * GEQO cycles. On the other hand, when we are marking a joinrel during GEQO, * we don't want the dummy path to clutter the main planning context. Upshot * is that the best solution is to explicitly make the dummy path in the same * context the given RelOptInfo is in. */ -static void +void mark_dummy_rel(RelOptInfo *rel) { MemoryContext oldcontext; @@ -993,9 +984,11 @@ mark_dummy_rel(RelOptInfo *rel) /* Evict any previously chosen paths */ rel->pathlist = NIL; + rel->partial_pathlist = NIL; /* Set up the dummy path */ - add_path(rel, (Path *) create_append_path(rel, NIL, NULL)); + add_path(rel, (Path *) create_append_path(NULL, rel, NIL, NIL, NULL, + 0, false, NIL, -1)); /* Set or update cheapest_total_path and related fields */ set_cheapest(rel); @@ -1003,6 +996,7 @@ mark_dummy_rel(RelOptInfo *rel) MemoryContextSwitchTo(oldcontext); } + /* * restriction_is_constant_false --- is a restrictlist just FALSE? * @@ -1012,10 +1006,13 @@ mark_dummy_rel(RelOptInfo *rel) * decide there's no match for an outer row, which is pretty stupid. So, * we need to detect the case. * - * If only_pushed_down is TRUE, then consider only pushed-down quals. + * If only_pushed_down is true, then consider only quals that are pushed-down + * from the point of view of the joinrel. */ static bool -restriction_is_constant_false(List *restrictlist, bool only_pushed_down) +restriction_is_constant_false(List *restrictlist, + RelOptInfo *joinrel, + bool only_pushed_down) { ListCell *lc; @@ -1027,10 +1024,9 @@ restriction_is_constant_false(List *restrictlist, bool only_pushed_down) */ foreach(lc, restrictlist) { - RestrictInfo *rinfo = (RestrictInfo *) lfirst(lc); + RestrictInfo *rinfo = lfirst_node(RestrictInfo, lc); - Assert(IsA(rinfo, RestrictInfo)); - if (only_pushed_down && !rinfo->is_pushed_down) + if (only_pushed_down && !RINFO_IS_PUSHED_DOWN(rinfo, joinrel->relids)) continue; if (rinfo->clause && IsA(rinfo->clause, Const)) @@ -1046,3 +1042,221 @@ restriction_is_constant_false(List *restrictlist, bool only_pushed_down) } return false; } + + +/* + * Set up tlist expressions for the childrel, and add EC members referencing + * the childrel. + */ +static void +update_child_rel_info(PlannerInfo *root, + RelOptInfo *rel, RelOptInfo *childrel) +{ + AppendRelInfo *appinfo = root->append_rel_array[childrel->relid]; + + /* Make child tlist expressions */ + childrel->reltarget->exprs = (List *) + adjust_appendrel_attrs(root, + (Node *) rel->reltarget->exprs, + 1, &appinfo); + + /* Make child entries in the EquivalenceClass as well */ + if (rel->has_eclass_joins || has_useful_pathkeys(root, rel)) + add_child_rel_equivalences(root, appinfo, rel, childrel); + childrel->has_eclass_joins = rel->has_eclass_joins; +} + +/* + * Construct the SpecialJoinInfo for a child-join by translating + * SpecialJoinInfo for the join between parents. left_relids and right_relids + * are the relids of left and right side of the join respectively. + */ +static SpecialJoinInfo * +build_child_join_sjinfo(PlannerInfo *root, SpecialJoinInfo *parent_sjinfo, + Relids left_relids, Relids right_relids) +{ + SpecialJoinInfo *sjinfo = makeNode(SpecialJoinInfo); + AppendRelInfo **left_appinfos; + int left_nappinfos; + AppendRelInfo **right_appinfos; + int right_nappinfos; + + memcpy(sjinfo, parent_sjinfo, sizeof(SpecialJoinInfo)); + left_appinfos = find_appinfos_by_relids(root, left_relids, + &left_nappinfos); + right_appinfos = find_appinfos_by_relids(root, right_relids, + &right_nappinfos); + + sjinfo->min_lefthand = adjust_child_relids(sjinfo->min_lefthand, + left_nappinfos, left_appinfos); + sjinfo->min_righthand = adjust_child_relids(sjinfo->min_righthand, + right_nappinfos, + right_appinfos); + sjinfo->syn_lefthand = adjust_child_relids(sjinfo->syn_lefthand, + left_nappinfos, left_appinfos); + sjinfo->syn_righthand = adjust_child_relids(sjinfo->syn_righthand, + right_nappinfos, + right_appinfos); + sjinfo->semi_rhs_exprs = (List *) adjust_appendrel_attrs(root, + (Node *) sjinfo->semi_rhs_exprs, + right_nappinfos, + right_appinfos); + + pfree(left_appinfos); + pfree(right_appinfos); + + return sjinfo; +} + +/* + * Assess whether join between given two partitioned relations can be broken + * down into joins between matching partitions; a technique called + * "partitionwise join" + * + * Partitionwise join is possible when a. Joining relations have same + * partitioning scheme b. There exists an equi-join between the partition keys + * of the two relations. + * + * Partitionwise join is planned as follows (details: optimizer/README.) + * + * 1. Create the RelOptInfos for joins between matching partitions i.e + * child-joins and add paths to them. + * + * 2. Construct Append or MergeAppend paths across the set of child joins. + * This second phase is implemented by generate_partitionwise_join_paths(). + * + * The RelOptInfo, SpecialJoinInfo and restrictlist for each child join are + * obtained by translating the respective parent join structures. + */ +static void +try_partitionwise_join(PlannerInfo *root, RelOptInfo *rel1, RelOptInfo *rel2, + RelOptInfo *joinrel, SpecialJoinInfo *parent_sjinfo, + List *parent_restrictlist) +{ + bool rel1_is_simple = IS_SIMPLE_REL(rel1); + bool rel2_is_simple = IS_SIMPLE_REL(rel2); + int nparts; + int cnt_parts; + + /* Guard against stack overflow due to overly deep partition hierarchy. */ + check_stack_depth(); + + /* Nothing to do, if the join relation is not partitioned. */ + if (!IS_PARTITIONED_REL(joinrel)) + return; + + /* The join relation should have consider_partitionwise_join set. */ + Assert(joinrel->consider_partitionwise_join); + + /* + * Since this join relation is partitioned, all the base relations + * participating in this join must be partitioned and so are all the + * intermediate join relations. + */ + Assert(IS_PARTITIONED_REL(rel1) && IS_PARTITIONED_REL(rel2)); + Assert(REL_HAS_ALL_PART_PROPS(rel1) && REL_HAS_ALL_PART_PROPS(rel2)); + + /* The joining relations should have consider_partitionwise_join set. */ + Assert(rel1->consider_partitionwise_join && + rel2->consider_partitionwise_join); + + /* + * The partition scheme of the join relation should match that of the + * joining relations. + */ + Assert(joinrel->part_scheme == rel1->part_scheme && + joinrel->part_scheme == rel2->part_scheme); + + /* + * Since we allow partitionwise join only when the partition bounds of the + * joining relations exactly match, the partition bounds of the join + * should match those of the joining relations. + */ + Assert(partition_bounds_equal(joinrel->part_scheme->partnatts, + joinrel->part_scheme->parttyplen, + joinrel->part_scheme->parttypbyval, + joinrel->boundinfo, rel1->boundinfo)); + Assert(partition_bounds_equal(joinrel->part_scheme->partnatts, + joinrel->part_scheme->parttyplen, + joinrel->part_scheme->parttypbyval, + joinrel->boundinfo, rel2->boundinfo)); + + nparts = joinrel->nparts; + + /* + * Create child-join relations for this partitioned join, if those don't + * exist. Add paths to child-joins for a pair of child relations + * corresponding to the given pair of parent relations. + */ + for (cnt_parts = 0; cnt_parts < nparts; cnt_parts++) + { + RelOptInfo *child_rel1 = rel1->part_rels[cnt_parts]; + RelOptInfo *child_rel2 = rel2->part_rels[cnt_parts]; + SpecialJoinInfo *child_sjinfo; + List *child_restrictlist; + RelOptInfo *child_joinrel; + Relids child_joinrelids; + AppendRelInfo **appinfos; + int nappinfos; + + /* + * If a child table has consider_partitionwise_join=false, it means + * that it's a dummy relation for which we skipped setting up tlist + * expressions and adding EC members in set_append_rel_size(), so do + * that now for use later. + */ + if (rel1_is_simple && !child_rel1->consider_partitionwise_join) + { + Assert(child_rel1->reloptkind == RELOPT_OTHER_MEMBER_REL); + Assert(IS_DUMMY_REL(child_rel1)); + update_child_rel_info(root, rel1, child_rel1); + child_rel1->consider_partitionwise_join = true; + } + if (rel2_is_simple && !child_rel2->consider_partitionwise_join) + { + Assert(child_rel2->reloptkind == RELOPT_OTHER_MEMBER_REL); + Assert(IS_DUMMY_REL(child_rel2)); + update_child_rel_info(root, rel2, child_rel2); + child_rel2->consider_partitionwise_join = true; + } + + /* We should never try to join two overlapping sets of rels. */ + Assert(!bms_overlap(child_rel1->relids, child_rel2->relids)); + child_joinrelids = bms_union(child_rel1->relids, child_rel2->relids); + appinfos = find_appinfos_by_relids(root, child_joinrelids, &nappinfos); + + /* + * Construct SpecialJoinInfo from parent join relations's + * SpecialJoinInfo. + */ + child_sjinfo = build_child_join_sjinfo(root, parent_sjinfo, + child_rel1->relids, + child_rel2->relids); + + /* + * Construct restrictions applicable to the child join from those + * applicable to the parent join. + */ + child_restrictlist = + (List *) adjust_appendrel_attrs(root, + (Node *) parent_restrictlist, + nappinfos, appinfos); + pfree(appinfos); + + child_joinrel = joinrel->part_rels[cnt_parts]; + if (!child_joinrel) + { + child_joinrel = build_child_join_rel(root, child_rel1, child_rel2, + joinrel, child_restrictlist, + child_sjinfo, + child_sjinfo->jointype); + joinrel->part_rels[cnt_parts] = child_joinrel; + } + + Assert(bms_equal(child_joinrel->relids, child_joinrelids)); + + populate_joinrel_with_paths(root, child_rel1, child_rel2, + child_joinrel, child_sjinfo, + child_restrictlist); + } +}