KingbaseES 数据库连接
一、数据准备：
create table student(
id int ,
s_name varchar(20),
t_id int
);
create table teacher(
id int ,
t_name varchar(20)
);
insert into student values (1,'zhangsan',3) , (2,'lisi',2) , (3,'wangwu',1), (4,'zhaoliu',4) ;
insert into teacher values(1, 'zhaolaoshi'),(2, 'qianlaoshi'),(3, 'sunlaoshi'),(6, 'lilaoshi');
以上为准备数据。
以下为测试准备SQL，不需要执行。
create index student_tid_idx on student(t_id);
create index student_id_idx on student(id);
create index teacher_id_idx on teacher(id);
reindex table student;
reindex table teacher;
set enable_seqscan to off;
set enable_hashjoin to off;
set enable_mergejoin to off;
drop index student_tid_idx;
drop index student_id_idx;
drop index teacher_id_idx;
二、数据库连接种类
1.自然连接
不需要指定连接列，也不需要使用on语句，默认按照相同列名进行筛选。
select * from student natural join teacher;
demo=# select * from student natural join teacher;
id | s_name | t_id | t_name
----+----------+------+------------
1 | zhangsan | 3 | zhaolaoshi
2 | lisi | 2 | qianlaoshi
3 | wangwu | 1 | sunlaoshi
(3 行记录)
2.内连接
（1）隐式语法
select * from student s,teacher t where s.t_id = t.id;
demo=# select * from student s,teacher t where s.t_id = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
(3 行记录)
（2）显示语法
写法1：
select * from student s inner join teacher t on s.t_id = t.id;
demo=# select * from student s inner join teacher t on s.t_id = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
(3 行记录)
写法2：
select * from student s join teacher t on s.t_id = t.id;
demo=# select * from student s join teacher t on s.t_id = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
(3 行记录)
3.外连接
（1）左外连接
写法1：
select * from student s left join teacher t on s.t_id = t.id;
demo=# select * from student s left join teacher t on s.t_id = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
4 | zhaoliu | 4 | |
(4 行记录)
写法2：
select * from student s left outer join teacher t on s.t_id = t.id;
demo=# select * from student s left outer join teacher t on s.t_id = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
4 | zhaoliu | 4 | |
(4 行记录)
写法3：
select * from student s,teacher t where s.t_id = t.id(+);
demo=# select * from student s,teacher t where s.t_id = t.id(+);
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
4 | zhaoliu | 4 | |
(4 行记录)
（2）右外连接
写法1：
select * from student s right join teacher t on s.t_id = t.id;
demo=# select * from student s right join teacher t on s.t_id = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
| | | 6 | lilaoshi
(4 行记录)
写法2：
select * from student s right outer join teacher t on s.t_id = t.id;
demo=# select * from student s right outer join teacher t on s.t_id = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
| | | 6 | lilaoshi
(4 行记录)
写法3：
select * from student s,teacher t where s.t_id(+) = t.id;
demo=# select * from student s,teacher t where s.t_id(+) = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
| | | 6 | lilaoshi
(4 行记录)
（3）全外连接
写法1：
select * from student s full join teacher t on s.t_id = t.id;
demo=# select * from student s full join teacher t on s.t_id = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
| | | 6 | lilaoshi
4 | zhaoliu | 4 | |
(5 行记录)
写法2：
select * from student s full outer join teacher t on s.t_id = t.id;
demo=# select * from student s full outer join teacher t on s.t_id = t.id;
id | s_name | t_id | id | t_name
----+----------+------+----+------------
3 | wangwu | 1 | 1 | zhaolaoshi
2 | lisi | 2 | 2 | qianlaoshi
1 | zhangsan | 3 | 3 | sunlaoshi
| | | 6 | lilaoshi
4 | zhaoliu | 4 | |
(5 行记录)
4.半连接
写法1：
select * from student s where exists (select * from teacher t where s.t_id = t.id) ;
demo=# select * from student s where exists (select * from teacher t where s.t_id = t.id) ;
id | s_name | t_id
----+----------+------
1 | zhangsan | 3
2 | lisi | 2
3 | wangwu | 1
(3 行记录)
写法2：
select * from student s where s.t_id in (select id from teacher t) ;
demo=# select * from student s where s.t_id in (select id from teacher t) ;
id | s_name | t_id
----+----------+------
1 | zhangsan | 3
2 | lisi | 2
3 | wangwu | 1
(3 行记录)
5.反半连接
写法1：
select * from student s where not exists (select * from teacher t where s.t_id = t.id) ;
demo=# select * from student s where not exists (select * from teacher t where s.t_id = t.id) ;
id | s_name | t_id
----+---------+------
4 | zhaoliu | 4
(1 行记录)
写法2：
select * from student s where s.t_id not in (select id from teacher t) ;
demo=# select * from student s where s.t_id not in (select id from teacher t) ;
id | s_name | t_id
----+---------+------
4 | zhaoliu | 4
(1 行记录)
三、数据库连接方法：
1.嵌套循环连接
（1）嵌套循环连接
嵌套循环连接是最基础的连接操作，任何连接条件都可以使用这种连接方式，但这种连接方式代价过高，因此PG中很少使用这种连接方式，更多的会使用以下几种嵌套循环连接的变种。
（2）物化嵌套循环连接
嵌套循环连接每读取一条左表(外表)中的元组时，都需要扫描右表(内表)中的所有元组，为每条外表记录做内表全扫描，这个过程代价高昂，
PG支持对内表做临时物化，并存储在内存或者临时文件中，在处理内表元组时，临时元组比缓冲区管理器更为高效，特别是内表比较小的时候，
可以全部载入内存时，我们将这种连接方式称之为物化嵌套循环连接
demo=# explain (analyze , buffers) select * from student s , teacher t ;
QUERY PLAN
-------------------------------------------------------------------------------------------------------------------
Nested Loop (cost=0.00..9714.75 rows=774000 width=128) (actual time=0.014..0.020 rows=16 loops=1)
Buffers: shared hit=2
-> Seq Scan on teacher t (cost=0.00..19.00 rows=900 width=62) (actual time=0.007..0.007 rows=4 loops=1)
Buffers: shared hit=1
-> Materialize (cost=0.00..22.90 rows=860 width=66) (actual time=0.001..0.002 rows=4 loops=4)
Buffers: shared hit=1
-> Seq Scan on student s (cost=0.00..18.60 rows=860 width=66) (actual time=0.002..0.003 rows=4 loops=1)
Buffers: shared hit=1
Planning Time: 0.045 ms
Execution Time: 0.037 ms
(10 行记录)
（3）外表索引嵌套循环连接
如果右表(内表)上有索引，且该索引能用于搜索满足连接条件的元组，那么计划器会为考虑使用索引进行直接搜索，以代替顺序扫描。
这种叫作索引嵌套循环连接
set enable_seqscan to off;
set enable_hashjoin to off;
set enable_mergejoin to off;
create index teacher_id_idx on teacher(id);
reindex table teacher;
demo=# explain (analyze , buffers) select * from student s , teacher t where s.t_id = t.id and t.id = 1;
QUERY PLAN
---------------------------------------------------------------------------------------------------------------------------------
Nested Loop (cost=10000000000.13..10000000009.21 rows=1 width=128) (actual time=0.013..0.014 rows=1 loops=1)
Buffers: shared hit=3
-> Seq Scan on student s (cost=10000000000.00..10000000001.05 rows=1 width=66) (actual time=0.006..0.007 rows=1 loops=1)
Filter: (t_id = 1)
Rows Removed by Filter: 3
Buffers: shared hit=1
-> Index Scan using teacher_id_idx on teacher t (cost=0.13..8.15 rows=1 width=62) (actual time=0.004..0.005 rows=1 loops=1)
Index Cond: (id = 1)
Buffers: shared hit=2
Planning Time: 0.119 ms
Execution Time: 0.031 ms
(11 行记录)
（4）外表索引扫描物化嵌套循环连接
在索引嵌套循环连接的定义基础上，将内表(右表)物化，注：内表需要有索引
set enable_seqscan to off;
create index teacher_id_idx on teacher(id);
reindex table teacher;
demo=# explain (analyze , buffers) select * from student s inner join teacher t on s.t_id = t.id and t.id <=6;
QUERY PLAN
---------------------------------------------------------------------------------------------------------------------------------------
Nested Loop (cost=10000000000.13..10000000009.25 rows=1 width=128) (actual time=0.016..0.021 rows=3 loops=1)
Join Filter: (s.t_id = t.id)
Rows Removed by Join Filter: 13
Buffers: shared hit=3
-> Seq Scan on student s (cost=10000000000.00..10000000001.04 rows=4 width=66) (actual time=0.005..0.005 rows=4 loops=1)
Buffers: shared hit=1
-> Materialize (cost=0.13..8.15 rows=1 width=62) (actual time=0.001..0.002 rows=4 loops=4)
Buffers: shared hit=2
-> Index Scan using teacher_id_idx on teacher t (cost=0.13..8.15 rows=1 width=62) (actual time=0.004..0.005 rows=4 loops=1)
Index Cond: (id <= 6)
Buffers: shared hit=2
Planning Time: 0.173 ms
Execution Time: 0.041 ms
(13 行记录)
（5）外表索引扫描索引嵌套循环连接
在索引嵌套循环连接的基础上，外表(左表)和内表(右表)上都有索引，且索引可以用于连接条件
set enable_seqscan to off;
set enable_hashjoin to off;
set enable_mergejoin to off;
create index student_tid_idx on student(t_id);
create index teacher_id_idx on teacher(id);
reindex table student;
reindex table teacher;
demo=# explain (analyze , buffers) select * from student s inner join teacher t on s.t_id = t.id and t.id <=6;
QUERY PLAN
----------------------------------------------------------------------------------------------------------------------------------
Nested Loop (cost=0.26..16.31 rows=1 width=128) (actual time=0.013..0.019 rows=3 loops=1)
Buffers: shared hit=9
-> Index Scan using teacher_id_idx on teacher t (cost=0.13..8.15 rows=1 width=62) (actual time=0.006..0.007 rows=4 loops=1)
Index Cond: (id <= 6)
Buffers: shared hit=2
-> Index Scan using student_tid_idx on student s (cost=0.13..8.15 rows=1 width=66) (actual time=0.001..0.001 rows=1 loops=4)
Index Cond: (t_id = t.id)
Buffers: shared hit=7
Planning Time: 0.242 ms
Execution Time: 0.038 ms
(10 行记录)
2.归并连接
（1）归并连接
内外表按照连接条件进行排序后，再进行连接，与嵌套循环连接不同的是，归并连接只适用于自然连接与等值连接
set enable_hashjoin to off;
set enable_mergejoin to on;
set enable_seqscan to on;
demo=# explain (analyze , buffers) select * from student s, teacher t where s.t_id = t.id and t.id > 1;
QUERY PLAN
----------------------------------------------------------------------------------------------------------------
Merge Join (cost=2.15..2.21 rows=3 width=29) (actual time=0.017..0.019 rows=2 loops=1)
Merge Cond: (s.t_id = t.id)
Buffers: shared hit=2
-> Sort (cost=1.08..1.09 rows=4 width=15) (actual time=0.008..0.009 rows=4 loops=1)
Sort Key: s.t_id
Sort Method: quicksort Memory: 25kB
Buffers: shared hit=1
-> Seq Scan on student s (cost=0.00..1.04 rows=4 width=15) (actual time=0.004..0.004 rows=4 loops=1)
Buffers: shared hit=1
-> Sort (cost=1.07..1.08 rows=3 width=14) (actual time=0.005..0.005 rows=3 loops=1)
Sort Key: t.id
Sort Method: quicksort Memory: 25kB
Buffers: shared hit=1
-> Seq Scan on teacher t (cost=0.00..1.05 rows=3 width=14) (actual time=0.003..0.003 rows=3 loops=1)
Filter: (id > 1)
Rows Removed by Filter: 1
Buffers: shared hit=1
Planning Time: 0.112 ms
Execution Time: 0.038 ms
(19 行记录)
（2）物化归并
在归并连接定义的基础上，将内表(右表)物化的归并
（3）外表索引扫描归并
归并连接定义的基础上，外表有索引，且索引适用于连接条件
set enable_seqscan to off;
create index teacher_id_idx on teacher(id);
reindex table teacher;
demo=# explain (analyze , buffers) select * from student s , teacher t where s.t_id = t.id and t.id>1;
QUERY PLAN
------------------------------------------------------------------------------------------------------------------------------------
Merge Join (cost=10000000001.21..10000000009.30 rows=3 width=29) (actual time=0.017..0.020 rows=2 loops=1)
Merge Cond: (s.t_id = t.id)
Buffers: shared hit=3
-> Sort (cost=10000000001.08..10000000001.09 rows=4 width=15) (actual time=0.009..0.009 rows=4 loops=1)
Sort Key: s.t_id
Sort Method: quicksort Memory: 25kB
Buffers: shared hit=1
-> Seq Scan on student s (cost=10000000000.00..10000000001.04 rows=4 width=15) (actual time=0.004..0.005 rows=4 loops=1)
Buffers: shared hit=1
-> Index Scan using teacher_id_idx on teacher t (cost=0.13..12.18 rows=3 width=14) (actual time=0.004..0.004 rows=3 loops=1)
Index Cond: (id > 1)
Buffers: shared hit=2
Planning Time: 0.122 ms
Execution Time: 0.042 ms
(14 行记录)
（4）外表索引扫描物化归并
在外表索引扫描归并连接的基础定义上，将内表物化的连接
（5）外表索引扫描索引归并连接
在外表索引扫描归并连接的基础上，内表也有索引，且索引可以用于连接条件
create index teacher_id_idx on teacher(id);
create index student_tid_idx on student(t_id);
reindex table teacher;
reindex table student;
demo=# explain (analyze , buffers) select * from student s , teacher t where s.t_id = t.id and t.id>1;
QUERY PLAN
-----------------------------------------------------------------------------------------------------------------------------------
Merge Join (cost=0.26..20.40 rows=3 width=29) (actual time=0.010..0.013 rows=2 loops=1)
Merge Cond: (s.t_id = t.id)
Buffers: shared hit=4
-> Index Scan using student_tid_idx on student s (cost=0.13..12.19 rows=4 width=15) (actual time=0.003..0.004 rows=4 loops=1)
Buffers: shared hit=2
-> Index Scan using teacher_id_idx on teacher t (cost=0.13..12.18 rows=3 width=14) (actual time=0.003..0.003 rows=3 loops=1)
Index Cond: (id > 1)
Buffers: shared hit=2
Planning Time: 0.269 ms
Execution Time: 0.033 ms
(10 行记录)
3.散列连接
（1）内存散列连接
与归并连接类似，散列连接只能适用于自然连接与等值连接，内存散列连接是在work_mem中处理，散列表区域被称为处理批次，一个处理批
次会有多少散列槽，被称为桶，桶的数量由nodeHash.c中定义的ExecChooseHashTableSize函数确定，且桶的数量总是2的整数次幂
set enable_hashjoin to on;
set enable_mergejoin to on;
set enable_seqscan to on;
demo=# explain (analyze , buffers) select * from student s, teacher t where s.t_id = t.id;
QUERY PLAN
----------------------------------------------------------------------------------------------------------------
Hash Join (cost=1.09..2.19 rows=4 width=29) (actual time=0.020..0.023 rows=3 loops=1)
Hash Cond: (s.t_id = t.id)
Buffers: shared hit=2
-> Seq Scan on student s (cost=0.00..1.04 rows=4 width=15) (actual time=0.007..0.008 rows=4 loops=1)
Buffers: shared hit=1
-> Hash (cost=1.04..1.04 rows=4 width=14) (actual time=0.006..0.007 rows=4 loops=1)
Buckets: 1024 Batches: 1 Memory Usage: 9kB
Buffers: shared hit=1
-> Seq Scan on teacher t (cost=0.00..1.04 rows=4 width=14) (actual time=0.002..0.003 rows=4 loops=1)
Buffers: shared hit=1
Planning Time: 0.067 ms
Execution Time: 0.042 ms
(12 行记录)
（2）带倾斜的混合散列连接
在内存散列定义的基础只上，如果内表的元组无法全部载入内存单个处理批次时，PG就用创建多个批次，只工作区中只分配一个批次，其它批次
都以临时文件的形式创建，属于某个批次的元组将通过临时元组存储功能写入相应的文件中
KingbaseES 数据库连接的更多相关文章
- KingbaseES 数据库连接断开问题排查思路
用户在使用数据库过程中,经常会发现如果会话空闲一段时间,会话有可能断开,需要重连.这个问题影响因素很多,包括数据库参数设置.操作系统参数.防火墙等.以下介绍KingbaseES针对该问题的排查思路. ...
- KingbaseES V8R3集群管理维护案例之---集群迁移单实例架构
案例说明: 在生产中,需要将KingbaseES V8R3集群转换为单实例架构,可以采用以下方式快速完成集群架构的迁移. 适用版本: KingbaseES V8R3 当前数据库版本: TEST=# s ...
- KingbaseES OOM 可能原因汇总
背景 不久前,遇到过一个问题.KingbaseES数据库主机系统触发OOM现象是数据库进程被KILL, 数据库进入crash然后restart(when restart_after_crash=on) ...
- KingbaseES ksqlrc文件介绍
ksqlrc文件作用 ksql在连接到数据库后但在接收正常的命令之前,会尝试读取并执行该文件中的命令,如果加上-X参数,则跳过该文件. 系统级的启动文件是ksqlrc,文件在安装好的KingbaseE ...
- KingbaseES R6 集群一键修改集群和数据库参数测试案例
案例说明: 集群环境修改集群或数据库参数,需要在每个node上都要修改,在每个节点而执行修改操作,容易出现漏改或节点上参数不一致等错误:在KingbaseES V8R6的集群中增加了,一键修改参数 ...
- KingbaseES R6 集群repmgr witness 手工配置案例
使用见证服务器: 见证服务器是一个正常的KingbaseES实例,不是流复制群集的一部分; 其目的是,如果发生故障转移情况,则提供证明它是主服务器本身不可用的证据,而不是例如在不同物理位置之间的网络分 ...
- jdbc访问KingbaseES数据库SocketTimeoutException Read timed out
jdbc访问KingbaseES数据库发生异常:SocketTimeoutException Read timed out 一.异常现象: 应用显示ERROR信息: Caused By: java.n ...
- Java数据库连接技术——JDBC
大家好,今天我们学习了Java如何连接数据库.之前学过.net语言的数据库操作,感觉就是一通百通,大同小异. JDBC是Java数据库连接技术的简称,提供连接各种常用数据库的能力. JDBC API ...
- .NET跨平台之旅：数据库连接字符串写法引发的问题
最近在一个ASP.NET Core站点中遇到一个奇怪问题.当用dotnet run命令启动站点后,开始的一段时间请求执行速度超慢,有时要超过20秒,有时甚至超过1分钟,日志中会记录这样的错误: Sys ...
- Entity Framework 6 Recipes 2nd Edition（12-3）译 -> 数据库连接日志
12-3. 数据库连接日志 问题 你想为每次与数据库的连接和断开记录日志 解决方案 EF为DbContext的连接公开了一个StateChange 事件.我们需要处理这个事件, 为每次与数据库的连接和 ...
随机推荐
- Go 项目的文件布局
转自 kcq 的 https://github.com/golang-standards/project-layout https://github.com/golang-standards/proj ...
- 麒麟系统开发笔记（十一）：在国产麒麟系统上使用gdb定位崩溃异常方法流程进阶定位代码行数及专项测试Demo
前言 上一篇,通过研究,可以定位到函数,本篇进一步优化,没有行数,程序较为复杂的时候,就无法定位,所以进一步定位. 本篇做了qBreakpad的研究,但是没有成功,过程也还是填出来,后来突然注意 ...
- 文件IO操作开发笔记（二）：使用Cpp的ofstream对磁盘文件存储进行性能测试以及测试工具
前言 在做到个别项目对日志要求较高,要求并行写入的数据较多,尽管写入数据的线程放在子线程,仍然会造成界面程序的假死(实际上Qt还是在跑,只是磁盘消耗超过瓶颈,造成假死(注意:控制台还能看到打印输出 ...
- Execl常用快捷操作
常用的操作 Ctrl+A 全选 Ctrl+Z 撤销 Ctrl+X 剪切 Ctrl+C 复制 Ctrl+V 粘贴 Ctrl+B 加粗 Ctrl+S 保存 Ctrl+F 查找 Ctrl+H 替换 Alt+ ...
- 【Azure Developer】use @azure/arm-monitor sdk 遇见 ManagedIdentityCredential authentication failed.(status code 500)
问题描述 在使用 @azure/arm-monitor sdk 创建 MonitorClient对象时候,遇见错误 ManagedIdentityCredential authentication f ...
- 16 Educational Codeforces Round 142 (Rated for Div. 2)C. Min Max Sort（递归、思维、dp）
C. Min Max Sort 很不错的一道题目,不过脑电波和出题人每对上,\(qwq.\) 正难则反. 我们考虑最后一步是怎么操作的. 最后一步一定是对\(1\)和\(n\)进行操作 那么上一步呢? ...
- Codeforces Round 113 (Div. 2)E. Tetrahedron（dp、递推）
目录 题面 链接 题意 题解 代码 总结 题面 链接 E. Tetrahedron 题意 从一个顶点出发走过路径长度为n回到出发点的方案总数 题解 考虑dp \(f[i][0|1|2|3]\):走了i ...
- k8s实战之MySQL单实例部署
前面我们学习了k8s入门系列文章,了解了k8s的一些基础概念以及怎么使用.本篇文章将进行一个小小的实战,使用k8s来部署单机版的mysql数据库,基本涵盖到前面讲到的Namespace.Pod.Dep ...
- husky 7.0.4 git hooks 前端 commit 钩子 git转svn [已解决]
husky 7 的安装,注意下版本 第一步 安装 cnpm install husky@7.0.4 --save-dev 第二步 在package.json script加入 "prepar ...
- 关于使用Kotlin开发SpringBoot项目使用@Transactional和@Autowired的报错问题
原文地址: 关于使用Kotlin开发SpringBoot项目使用@Transactional和@Autowired的报错问题 - Stars-One的杂货小窝 问题描述 最近在开发一个订单模块,需要出 ...