PostgreSQL 源码解读（25）- 查询语句#10（查询优化概览）

本节简单介绍了PG执行查询语句中优化器部分（Optimizer）的相关函数和数据结构总体说明。查询优化包括查询逻辑优化和查询物理优化，查询逻辑优化是指使用关系代数中的等价规则，通过选择下推、投影下推、连接交换等方法对SQL语句进行优化;查询物理优化是指通过CBO对各种物理访问数据的方法进行评估，得出最优的执行计划。

一、总体说明

下面是PG源码目录(/src/backend/optimizer)中的README文件对优化器相关函数和数据结构的总体说明:

Optimizer Functions
-------------------

The primary entry point is planner().

planner()//优化器主入口函数
set up for recursive handling of subqueries//为子查询配置处理器(递归方式)
-subquery_planner()//调用(子)查询优化函数
 pull up sublinks and subqueries from rangetable, if possible//可以的话,上拉子链接和子查询
 canonicalize qual//表达式规范化
     Attempt to simplify WHERE clause to the most useful form; this includes
     flattening nested AND/ORs and detecting clauses that are duplicated in
     different branches of an OR.//简化WHERE语句
 simplify constant expressions//简化常量表达式
 process sublinks//处理子链接
 convert Vars of outer query levels into Params//转换外查询的Vars变量到Params中
--grouping_planner()//
  preprocess target list for non-SELECT queries//预处理非SELECT语句的投影列
  handle UNION/INTERSECT/EXCEPT, GROUP BY, HAVING, aggregates,//处理集合操作/聚集函数/排序等
    ORDER BY, DISTINCT, LIMIT
--query_planner()//
   make list of base relations used in query//构造查询中的基表链表
   split up the qual into restrictions (a=1) and joins (b=c)//拆分表达式为限制条件和连接
   find qual clauses that enable merge and hash joins//查找可以让Merge和Hash连接生效的表达式
----make_one_rel()//
     set_base_rel_pathlists()//设置基表路径链表
      find seqscan and all index paths for each base relation//遍历每个基表,寻找顺序扫描和所有可能的索引扫描路径
      find selectivity of columns used in joins//查找连接中使用的列的选择性
     make_rel_from_joinlist()//通过join链表构造Relation
      hand off join subproblems to a plugin, GEQO, or standard_join_search()//
-----standard_join_search()//标准的连接搜索函数
      call join_search_one_level() for each level of join tree needed//每一个join tree调用join_search_one_level
      join_search_one_level():
        For each joinrel of the prior level, do make_rels_by_clause_joins()//对于上一层的每一个joinrel,执行make_rels_by_clause_joins
        if it has join clauses, or make_rels_by_clauseless_joins() if not.
        Also generate "bushy plan" joins between joinrels of lower levels.
      Back at standard_join_search(), generate gather paths if needed for//回到standard_join_search函数,需要的话,收集相关的路径并应用set_cheapest函数获取代价最小的路径
      each newly constructed joinrel, then apply set_cheapest() to extract
      the cheapest path for it.
      Loop back if this was not the top join level.//如果不是最顶层连接,循环
  Back at grouping_planner://回到grouping_planner函数
  do grouping (GROUP BY) and aggregation//处理分组和聚集
  do window functions//处理窗口函数
  make unique (DISTINCT)//处理唯一性
  do sorting (ORDER BY)//处理排序
  do limit (LIMIT/OFFSET)//处理Limit
Back at planner()://回到planner函数
convert finished Path tree into a Plan tree//转换最终的路径树到计划树
do final cleanup after planning//收尾工作


Optimizer Data Structures
-------------------------

PlannerGlobal   - global information for a single planner invocation//全局优化信息

PlannerInfo     - information for planning a particular Query (we make//某个Planner的优化信息
                  a separate PlannerInfo node for each sub-Query)

RelOptInfo      - a relation or joined relations//某个Relation(包括连接)的优化信息

 RestrictInfo   - WHERE clauses, like "x = 3" or "y = z"//限制条件
                  (note the same structure is used for restriction and
                   join clauses)

 Path           - every way to generate a RelOptInfo(sequential,index,joins)//构造该关系(注意:中间结果也是关系的一种)的路径
  SeqScan       - represents a sequential scan plan
  IndexPath     - index scan
  BitmapHeapPath - top of a bitmapped index scan
  TidPath       - scan by CTID
  SubqueryScanPath - scan a subquery-in-FROM
  ForeignPath   - scan a foreign table, foreign join or foreign upper-relation
  CustomPath    - for custom scan providers
  AppendPath    - append multiple subpaths together
  MergeAppendPath - merge multiple subpaths, preserving their common sort order
  ResultPath    - a childless Result plan node (used for FROM-less SELECT)
  MaterialPath  - a Material plan node
  UniquePath    - remove duplicate rows (either by hashing or sorting)
  GatherPath    - collect the results of parallel workers
  GatherMergePath - collect parallel results, preserving their common sort order
  ProjectionPath - a Result plan node with child (used for projection)
  ProjectSetPath - a ProjectSet plan node applied to some sub-path
  SortPath      - a Sort plan node applied to some sub-path
  GroupPath     - a Group plan node applied to some sub-path
  UpperUniquePath - a Unique plan node applied to some sub-path
  AggPath       - an Agg plan node applied to some sub-path
  GroupingSetsPath - an Agg plan node used to implement GROUPING SETS
  MinMaxAggPath - a Result plan node with subplans performing MIN/MAX
  WindowAggPath - a WindowAgg plan node applied to some sub-path
  SetOpPath     - a SetOp plan node applied to some sub-path
  RecursiveUnionPath - a RecursiveUnion plan node applied to two sub-paths
  LockRowsPath  - a LockRows plan node applied to some sub-path
  ModifyTablePath - a ModifyTable plan node applied to some sub-path(s)
  LimitPath     - a Limit plan node applied to some sub-path
  NestPath      - nested-loop joins
  MergePath     - merge joins
  HashPath      - hash joins

 EquivalenceClass - a data structure representing a set of values known equal//等价类

 PathKey        - a data structure representing the sort ordering of a path//排序键

下一节开始将根据总体说明中的函数逐个进行分析解读.

二、小结

1、优化器函数总览：大体介绍了优化器函数的调用过程等信息；
2、数据结构：优化器相关的数据结构，如PlannerInfo等。