服务器之家

服务器之家 > 正文

postgresql insert into select无法使用并行查询的解决

时间:2021-03-15 18:00     来源/作者:瀚高PG实验室

本文信息基于PG13.1。

从PG9.6开始支持并行查询。PG11开始支持CREATE TABLE … AS、SELECT INTO以及CREATE MATERIALIZED VIEW的并行查询。

先说结论:

换用create table as 或者select into或者导入导出。

首先跟踪如下查询语句的执行计划:

?
1
select count(*) from test t1,test1 t2 where t1.id = t2.id ;
?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
postgres=# explain analyze select count(*) from test t1,test1 t2 where t1.id = t2.id ;
                                    QUERY PLAN                                   
--------------------------------------------------------------------------------------------------------------------------------------------------------
 Finalize Aggregate (cost=34244.16..34244.17 rows=1 width=8) (actual time=683.246..715.324 rows=1 loops=1)
  -> Gather (cost=34243.95..34244.16 rows=2 width=8) (actual time=681.474..715.311 rows=3 loops=1)
     Workers Planned: 2
     Workers Launched: 2
     -> Partial Aggregate (cost=33243.95..33243.96 rows=1 width=8) (actual time=674.689..675.285 rows=1 loops=3)
        -> Parallel Hash Join (cost=15428.00..32202.28 rows=416667 width=0) (actual time=447.799..645.689 rows=333333 loops=3)
           Hash Cond: (t1.id = t2.id)
           -> Parallel Seq Scan on test t1 (cost=0.00..8591.67 rows=416667 width=4) (actual time=0.025..74.010 rows=333333 loops=3)
           -> Parallel Hash (cost=8591.67..8591.67 rows=416667 width=4) (actual time=260.052..260.053 rows=333333 loops=3)
              Buckets: 131072 Batches: 16 Memory Usage: 3520kB
              -> Parallel Seq Scan on test1 t2 (cost=0.00..8591.67 rows=416667 width=4) (actual time=0.032..104.804 rows=333333 loops=3)
 Planning Time: 0.420 ms
 Execution Time: 715.447 ms
(13 rows)

可以看到走了两个Workers。

下边看一下insert into select:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
postgres=# explain analyze insert into va select count(*) from test t1,test1 t2 where t1.id = t2.id ;    
                                  QUERY PLAN                                 
--------------------------------------------------------------------------------------------------------------------------------------------------
 Insert on va (cost=73228.00..73228.02 rows=1 width=4) (actual time=3744.179..3744.187 rows=0 loops=1)
  -> Subquery Scan on "*SELECT*" (cost=73228.00..73228.02 rows=1 width=4) (actual time=3743.343..3743.352 rows=1 loops=1)
     -> Aggregate (cost=73228.00..73228.01 rows=1 width=8) (actual time=3743.247..3743.254 rows=1 loops=1)
        -> Hash Join (cost=30832.00..70728.00 rows=1000000 width=0) (actual time=1092.295..3511.301 rows=1000000 loops=1)
           Hash Cond: (t1.id = t2.id)
           -> Seq Scan on test t1 (cost=0.00..14425.00 rows=1000000 width=4) (actual time=0.030..421.537 rows=1000000 loops=1)
           -> Hash (cost=14425.00..14425.00 rows=1000000 width=4) (actual time=1090.078..1090.081 rows=1000000 loops=1)
              Buckets: 131072 Batches: 16 Memory Usage: 3227kB
              -> Seq Scan on test1 t2 (cost=0.00..14425.00 rows=1000000 width=4) (actual time=0.021..422.768 rows=1000000 loops=1)
 Planning Time: 0.511 ms
 Execution Time: 3745.633 ms
(11 rows)

可以看到并没有Workers的指示,没有启用并行查询。

即使开启强制并行,也无法走并行查询。

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
postgres=# set force_parallel_mode =on;
SET
postgres=# explain analyze insert into va select count(*) from test t1,test1 t2 where t1.id = t2.id ;
                                  QUERY PLAN                                 
--------------------------------------------------------------------------------------------------------------------------------------------------
 Insert on va (cost=73228.00..73228.02 rows=1 width=4) (actual time=3825.042..3825.049 rows=0 loops=1)
  -> Subquery Scan on "*SELECT*" (cost=73228.00..73228.02 rows=1 width=4) (actual time=3824.976..3824.984 rows=1 loops=1)
     -> Aggregate (cost=73228.00..73228.01 rows=1 width=8) (actual time=3824.972..3824.978 rows=1 loops=1)
        -> Hash Join (cost=30832.00..70728.00 rows=1000000 width=0) (actual time=1073.587..3599.402 rows=1000000 loops=1)
           Hash Cond: (t1.id = t2.id)
           -> Seq Scan on test t1 (cost=0.00..14425.00 rows=1000000 width=4) (actual time=0.034..414.965 rows=1000000 loops=1)
           -> Hash (cost=14425.00..14425.00 rows=1000000 width=4) (actual time=1072.441..1072.443 rows=1000000 loops=1)
              Buckets: 131072 Batches: 16 Memory Usage: 3227kB
              -> Seq Scan on test1 t2 (cost=0.00..14425.00 rows=1000000 width=4) (actual time=0.022..400.624 rows=1000000 loops=1)
 Planning Time: 0.577 ms
 Execution Time: 3825.923 ms
(11 rows)

原因在官方文档有写:

The query writes any data or locks any database rows. If a query contains a data-modifying operation either at the top level or within a CTE, no parallel plans for that query will be generated. As an exception, the commands CREATE TABLE … AS, SELECT INTO, and CREATE MATERIALIZED VIEW which create a new table and populate it can use a parallel plan.

解决方案有如下三种:

 

1.select into

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
postgres=# explain analyze select count(*) into vaa from test t1,test1 t2 where t1.id = t2.id ;
                                    QUERY PLAN                                   
--------------------------------------------------------------------------------------------------------------------------------------------------------
 Finalize Aggregate (cost=34244.16..34244.17 rows=1 width=8) (actual time=742.736..774.923 rows=1 loops=1)
  -> Gather (cost=34243.95..34244.16 rows=2 width=8) (actual time=740.223..774.907 rows=3 loops=1)
     Workers Planned: 2
     Workers Launched: 2
     -> Partial Aggregate (cost=33243.95..33243.96 rows=1 width=8) (actual time=731.408..731.413 rows=1 loops=3)
        -> Parallel Hash Join (cost=15428.00..32202.28 rows=416667 width=0) (actual time=489.880..700.830 rows=333333 loops=3)
           Hash Cond: (t1.id = t2.id)
           -> Parallel Seq Scan on test t1 (cost=0.00..8591.67 rows=416667 width=4) (actual time=0.033..87.479 rows=333333 loops=3)
           -> Parallel Hash (cost=8591.67..8591.67 rows=416667 width=4) (actual time=266.839..266.840 rows=333333 loops=3)
              Buckets: 131072 Batches: 16 Memory Usage: 3520kB
              -> Parallel Seq Scan on test1 t2 (cost=0.00..8591.67 rows=416667 width=4) (actual time=0.058..106.874 rows=333333 loops=3)
 Planning Time: 0.319 ms
 Execution Time: 783.300 ms
(13 rows)

2.create table as

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
postgres=# explain analyze create table vb as select count(*) from test t1,test1 t2 where t1.id = t2.id ;
                                   QUERY PLAN                                   
-------------------------------------------------------------------------------------------------------------------------------------------------------
 Finalize Aggregate (cost=34244.16..34244.17 rows=1 width=8) (actual time=540.120..563.733 rows=1 loops=1)
  -> Gather (cost=34243.95..34244.16 rows=2 width=8) (actual time=537.982..563.720 rows=3 loops=1)
     Workers Planned: 2
     Workers Launched: 2
     -> Partial Aggregate (cost=33243.95..33243.96 rows=1 width=8) (actual time=526.602..527.136 rows=1 loops=3)
        -> Parallel Hash Join (cost=15428.00..32202.28 rows=416667 width=0) (actual time=334.532..502.793 rows=333333 loops=3)
           Hash Cond: (t1.id = t2.id)
           -> Parallel Seq Scan on test t1 (cost=0.00..8591.67 rows=416667 width=4) (actual time=0.018..57.819 rows=333333 loops=3)
           -> Parallel Hash (cost=8591.67..8591.67 rows=416667 width=4) (actual time=189.502..189.503 rows=333333 loops=3)
              Buckets: 131072 Batches: 16 Memory Usage: 3520kB
              -> Parallel Seq Scan on test1 t2 (cost=0.00..8591.67 rows=416667 width=4) (actual time=0.023..77.786 rows=333333 loops=3)
 Planning Time: 0.189 ms
 Execution Time: 565.448 ms
(13 rows)

3.或者通过导入导出的方式,例如:

?
1
2
psql -h localhost -d postgres -U postgres -c "select count(*) from test t1,test1 t2 where t1.id = t2.id " -o result.csv -A -t -F ","
psql -h localhost -d postgres -U postgres -c "COPY va FROM 'result.csv' WITH (FORMAT CSV, DELIMITER ',', HEADER FALSE, ENCODING 'windows-1252')"

一些场景下也会比非并行快。

到此这篇关于postgresql insert into select无法使用并行查询的解决的文章就介绍到这了,更多相关postgresql insert into select并行查询内容请搜索服务器之家以前的文章或继续浏览下面的相关文章希望大家以后多多支持服务器之家!

原文链接:https://blog.csdn.net/pg_hgdb/article/details/112297250

相关文章

热门资讯

2020微信伤感网名听哭了 让对方看到心疼的伤感网名大全
2020微信伤感网名听哭了 让对方看到心疼的伤感网名大全 2019-12-26
yue是什么意思 网络流行语yue了是什么梗
yue是什么意思 网络流行语yue了是什么梗 2020-10-11
背刺什么意思 网络词语背刺是什么梗
背刺什么意思 网络词语背刺是什么梗 2020-05-22
Intellij idea2020永久破解,亲测可用!!!
Intellij idea2020永久破解,亲测可用!!! 2020-07-29
苹果12mini价格表官网报价 iPhone12mini全版本价格汇总
苹果12mini价格表官网报价 iPhone12mini全版本价格汇总 2020-11-13
返回顶部