共計 5598 個字符，預計需要花費 14 分鐘才能閱讀完成。

丸趣 TV 小編給大家分享一下 PostgreSQL 中 Review PG 的 Optimizer 機制如何優化函數，相信大部分人都還不怎么了解，因此分享這篇文章給大家參考一下，希望大家閱讀完這篇文章后大有收獲，下面讓我們一起去了解一下吧！

一、Optimizer Functions

Optimizer Functions- 查詢優化函數

The primary entry point is planner().
planner() // 主入口
set up for recursive handling of subqueries
-subquery_planner()//planner- subquery_planner
pull up sublinks and subqueries from rangetable, if possible
canonicalize qual
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.
simplify constant expressions
process sublinks
convert Vars of outer query levels into Params
–grouping_planner()//planner- subquery_planner- grouping_planner
preprocess target list for non-SELECT queries
handle UNION/INTERSECT/EXCEPT, GROUP BY, HAVING, aggregates,
ORDER BY, DISTINCT, LIMIT
—query_planner()//subquery_planner- grouping_planner- query_planner
make list of base relations used in query
split up the qual into restrictions (a=1) and joins (b=c)
find qual clauses that enable merge and hash joins
—-make_one_rel()//…grouping_planner- query_planner- make_one_rel
set_base_rel_pathlists() // 為每一個 RelOptInfo 生成訪問路徑
find seqscan and all index paths for each base relation
find selectivity of columns used in joins
make_rel_from_joinlist() // 使用遺傳算法或動態規劃算法構造連接路徑
hand off join subproblems to a plugin, GEQO, or standard_join_search()
—–standard_join_search()// 這是動態規劃算法
call join_search_one_level() for each level of join tree needed
join_search_one_level():
For each joinrel of the prior level, do 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
each newly constructed joinrel, then apply set_cheapest() to extract
the cheapest path for it.
Loop back if this wasn t the top join level.
Back at grouping_planner:
do grouping (GROUP BY) and aggregation// 在最高層處理分組 / 聚集 / 唯一過濾 / 排序 / 控制輸出元組數目等
do window functions
make unique (DISTINCT)
do sorting (ORDER BY)
do limit (LIMIT/OFFSET)
Back at planner():
convert finished Path tree into a Plan tree
do final cleanup after planning

二、Optimizer Data Structures

Optimizer Data Structures
數據結構

PlannerGlobal   – global information for a single planner invocation
PlannerInfo     – information for planning a particular Query (we make
a separate PlannerInfo node for each sub-Query)
RelOptInfo      – a relation or joined relations
RestrictInfo   – WHERE clauses, like x = 3 or y = z
(note the same structure is used for restriction and
join clauses)
Path           – every way to generate a RelOptInfo(sequential,index,joins)
SeqScan       – represents a sequential scan plan // 順序掃描
IndexPath     – index scan // 索引掃描
BitmapHeapPath – top of a bitmapped index scan // 位圖索引掃描
TidPath       – scan by CTID //CTID 掃描
SubqueryScanPath – scan a subquery-in-FROM //FROM 子句中的子查詢掃描
ForeignPath   – scan a foreign table, foreign join or foreign upper-relation //FDW
CustomPath    – for custom scan providers // 定制化掃描
AppendPath    – append multiple subpaths together // 多個子路徑 APPEND, 常見于集合操作
MergeAppendPath – merge multiple subpaths, preserving their common sort order // 保持順序的 APPEND
ResultPath    – a childless Result plan node (used for FROM-less SELECT)// 結果路徑 (如 SELECT 2+2)
MaterialPath  – a Material plan node // 物化路徑
UniquePath    – remove duplicate rows (either by hashing or sorting) // 去除重復行路徑
GatherPath    – collect the results of parallel workers // 并行
GatherMergePath – collect parallel results, preserving their common sort order // 并行, 保持順序
ProjectionPath – a Result plan node with child (used for projection) // 投影
ProjectSetPath – a ProjectSet plan node applied to some sub-path // 投影 (應用于子路徑上)
SortPath      – a Sort plan node applied to some sub-path // 排序
GroupPath     – a Group plan node applied to some sub-path // 分組
UpperUniquePath – a Unique plan node applied to some sub-path // 應用于子路徑的 Unique Plan
AggPath       – an Agg plan node applied to some sub-path // 應用于子路徑的聚集
GroupingSetsPath – an Agg plan node used to implement GROUPING SETS // 分組集合
MinMaxAggPath – a Result plan node with subplans performing MIN/MAX // 最大最小
WindowAggPath – a WindowAgg plan node applied to some sub-path // 應用于子路徑的窗口函數
SetOpPath     – a SetOp plan node applied to some sub-path // 應用于子路徑的集合操作
RecursiveUnionPath – a RecursiveUnion plan node applied to two sub-paths // 遞歸 UNION
LockRowsPath  – a LockRows plan node applied to some sub-path // 應用于子路徑的的 LockRows
ModifyTablePath – a ModifyTable plan node applied to some sub-path(s) // 應用于子路徑的數據表更新 (如 INSERT/UPDATE 操作等)
LimitPath     – a Limit plan node applied to some sub-path// 應用于子路徑的 LIMIT
NestPath      – nested-loop joins// 嵌套循環連接
MergePath     – merge joins//Merge Join
HashPath      – hash joins//Hash Join
EquivalenceClass – a data structure representing a set of values known equal
PathKey        – a data structure representing the sort ordering of a path

The optimizer spends a good deal of its time worrying about the ordering
of the tuples returned by a path.  The reason this is useful is that by
knowing the sort ordering of a path, we may be able to use that path as
the left or right input of a mergejoin and avoid an explicit sort step.
Nestloops and hash joins don t really care what the order of their inputs
is, but mergejoin needs suitably ordered inputs.  Therefore, all paths
generated during the optimization process are marked with their sort order
(to the extent that it is known) for possible use by a higher-level merge.

優化器在元組的排序上面花費了不少時間, 原因是為了在 Merge Join 時避免專門的排序步驟.

It is also possible to avoid an explicit sort step to implement a user s
ORDER BY clause if the final path has the right ordering already, so the
sort ordering is of interest even at the top level.  grouping_planner() will
look for the cheapest path with a sort order matching the desired order,
then compare its cost to the cost of using the cheapest-overall path and
doing an explicit sort on that.
When we are generating paths for a particular RelOptInfo, we discard a path
if it is more expensive than another known path that has the same or better
sort order.  We will never discard a path that is the only known way to
achieve a given sort order (without an explicit sort, that is).  In this
way, the next level up will have the maximum freedom to build mergejoins
without sorting, since it can pick from any of the paths retained for its
inputs.

以上是“PostgreSQL 中 Review PG 的 Optimizer 機制如何優化函數”這篇文章的所有內容，感謝各位的閱讀！相信大家都有了一定的了解，希望分享的內容對大家有所幫助，如果還想學習更多知識，歡迎關注丸趣 TV 行業資訊頻道！