[TOC]
# 分布式计划的生成
OceanBase 数据库的优化器会分为以下两大阶段来生成分布式的执行计划。
1\. 第一阶段：不考虑数据的物理分布，生成所有基于本地关系优化的最优执行计划。在本地计划生成后，优化器会检查数据是否访问了多个分区，或者是否访问的是本地单分区表但是用户使用 HINT 强制采用了并行查询执行。
2\. 第二阶段：生成分布式计划。根据执行计划树，在需要进行数据重分布的地方，插入 EXCHANGE 节点，从而将原先的本地计划树变成分布式执行计划。
## 分布式执行计划的算子
生成分布式计划的过程就是在原始计划树上寻找恰当位置插入 EXCHANGE 算子的过程，在自顶向下遍历计划树的时候，需要根据相应算子的数据处理情况以及输入算子的数据分区情况，来决定是否需要插入 EXCHANGE 算子。
如下示例为最简单的单表扫描：
~~~
obclient>CREATE TABLE t1 (v1 INT, v2 INT) PARTITION BY HASH(v1) PARTITIONS 5;
Query OK, 0 rows affected (0.12 sec)
obclient>EXPLAIN SELECT * FROM t1\G;
*************************** 1. row ***************************
Query Plan:
==============================================
|ID|OPERATOR |NAME |EST. ROWS|COST |
------------------------------------------------------
|0 |PX COORDINATOR | |500000 |545109|
|1 | EXCHANGE OUT DISTR |:EX10000|500000 |320292|
|2 | PX PARTITION ITERATOR| |500000 |320292|
|3 | TABLE SCAN |T1 |500000 |320292|
======================================================
Outputs & filters:
-------------------------------------
0 - output([T1.V1], [T1.V2]), filter(nil)
1 - output([T1.V1], [T1.V2]), filter(nil), dop=1
2 - output([T1.V1], [T1.V2]), filter(nil)
3 - output([T1.V1], [T1.V2]), filter(nil),
access([T1.V1], [T1.V2]), partitions(p[0-4])
~~~
当 t1 是一个分区表，可以在 TABLE SCAN 上插入配对的 EXCHANGE 算子，从而将 TABLE SCAN 和 EXCHANGE OUT 封装成一个 job，可以用于并行的执行。
**单输入可下压算子**
单输入可下压算子主要包括 AGGREGATION、SORT、GROUP BY 和 LIMIT 算子等，除了 LIMIT 算子以外，其余所列举的算子都会有一个操作的键，如果操作的键和输入数据的数据分布是一致的，则可以做一阶段聚合操作，也即 Partition Wise Aggregation。如果操作的键和输入数据的数据分布是不一致的，则需要做两阶段聚合操作，聚合算子需要做下压操作。
一阶段聚合操作如下例所示：
~~~
obclient>CREATE TABLE t2 (v1 INT, v2 INT) PARTITION BY HASH(v1) PARTITIONS 4;
Query OK, 0 rows affected (0.12 sec)
obclient>EXPLAIN SELECT SUM(v1) FROM t2 GROUP BY v1\G;
*************************** 1. row ***************************
Query Plan:
| ======================================================
|ID|OPERATOR |NAME |EST. ROWS|COST |
------------------------------------------------------
|0 |PX COORDINATOR | |101 |357302|
|1 | EXCHANGE OUT DISTR |:EX10000|101 |357297|
|2 | PX PARTITION ITERATOR| |101 |357297|
|3 | MERGE GROUP BY | |101 |357297|
|4 | TABLE SCAN |t2 |400000 |247403|
======================================================
Outputs & filters:
-------------------------------------
0 - output([T_FUN_SUM(t2.v1)]), filter(nil)
1 - output([T_FUN_SUM(t2.v1)]), filter(nil), dop=1
2 - output([T_FUN_SUM(t2.v1)]), filter(nil)
3 - output([T_FUN_SUM(t2.v1)]), filter(nil),
group([t2.v1]), agg_func([T_FUN_SUM(t2.v1)])
4 - output([t2.v1]), filter(nil),
access([t2.v1]), partitions(p[0-3])
~~~
二阶段聚合操作如下例所示：
~~~
| ============================================================
|ID|OPERATOR |NAME |EST. ROWS|COST |
------------------------------------------------------------
|0 |PX COORDINATOR | |101 |561383|
|1 | EXCHANGE OUT DISTR |:EX10001|101 |561374|
|2 | HASH GROUP BY | |101 |561374|
|3 | EXCHANGE IN DISTR | |101 |408805|
|4 | EXCHANGE OUT DISTR (HASH)|:EX10000|101 |408795|
|5 | HASH GROUP BY | |101 |408795|
|6 | PX PARTITION ITERATOR | |400000 |256226|
|7 | TABLE SCAN |t2 |400000 |256226|
============================================================
Outputs & filters:
-------------------------------------
0 - output([T_FUN_SUM(T_FUN_SUM(t2.v1))]), filter(nil)
1 - output([T_FUN_SUM(T_FUN_SUM(t2.v1))]), filter(nil), dop=1
2 - output([T_FUN_SUM(T_FUN_SUM(t2.v1))]), filter(nil),
group([t2.v2]), agg_func([T_FUN_SUM(T_FUN_SUM(t2.v1))])
3 - output([t2.v2], [T_FUN_SUM(t2.v1)]), filter(nil)
4 - (#keys=1, [t2.v2]), output([t2.v2], [T_FUN_SUM(t2.v1)]), filter(nil), dop=1
5 - output([t2.v2], [T_FUN_SUM(t2.v1)]), filter(nil),
group([t2.v2]), agg_func([T_FUN_SUM(t2.v1)])
6 - output([t2.v1], [t2.v2]), filter(nil)
7 - output([t2.v1], [t2.v2]), filter(nil),
access([t2.v1], [t2.v2]), partitions(p[0-3])
~~~
**二元输入算子**
二元输入算子主要考虑 JOIN 算子的情况。对于 JOIN 算子来说，主要基于规则来生成分布式执行计划和选择数据重分布方法。JOIN 算子主要有以下三种联接方式：
* Partition-Wise Join
当左右表都是分区表且分区方式相同，物理分布一样，并且 JOIN 的联接条件为分区键时，可以使用以分区为单位的联接方法。如下例所示：
~~~
obclient>CREATE TABLE t3 (v1 INT, v2 INT) PARTITION BY HASH(v1) PARTITIONS 4;
Query OK, 0 rows affected (0.12 sec)
obclient>EXPLAIN SELECT * FROM t2, t3 WHERE t2.v1 = t3.v1\G;
*************************** 1. row ***************************
Query Plan:
===========================================================
|ID|OPERATOR |NAME |EST. ROWS |COST |
|0 |PX COORDINATOR | |1568160000|1227554264|
|1 | EXCHANGE OUT DISTR |:EX10000|1568160000|930670004 |
|2 | PX PARTITION ITERATOR| |1568160000|930670004 |
|3 | MERGE JOIN | |1568160000|930670004 |
|4 | TABLE SCAN |t2 |400000 |256226 |
|5 | TABLE SCAN |t3 |400000 |256226 |
===========================================================
Outputs & filters:
-------------------------------------
0 - output([t2.v1], [t2.v2], [t3.v1], [t3.v2]), filter(nil)
1 - output([t2.v1], [t2.v2], [t3.v1], [t3.v2]), filter(nil), dop=1
2 - output([t2.v1], [t2.v2], [t3.v1], [t3.v2]), filter(nil)
3 - output([t2.v1], [t2.v2], [t3.v1], [t3.v2]), filter(nil),
equal_conds([t2.v1 = t3.v1]), other_conds(nil)
4 - output([t2.v1], [t2.v2]), filter(nil),
access([t2.v1], [t2.v2]), partitions(p[0-3])
5 - output([t3.v1], [t3.v2]), filter(nil),
access([t3.v1], [t3.v2]), partitions(p[0-3])
~~~
* Partial Partition-Wise Join
当左右表中一个表为分区表，另一个表为非分区表，或者两者皆为分区表但是联接键仅和其中一个分区表的分区键相同的情况下，会以该分区表的分区分布为基准，重新分布另一个表的数据。如下例所示：
~~~
obclient>CREATE TABLE t4 (v1 INT, v2 INT) PARTITION BY HASH(v1) PARTITIONS 3;
Query OK, 0 rows affected (0.12 sec)
obclient>EXPLAIN SELECT * FROM t4, t2 WHERE t2.v1 = t4.v1\G;
*************************** 1. row ***************************
Query Plan:
===========================================================
|ID|OPERATOR |NAME |EST. ROWS|COST |
-----------------------------------------------------------
|0 |PX COORDINATOR | |11880 |17658|
|1 | EXCHANGE OUT DISTR |:EX10001|11880 |15409|
|2 | NESTED-LOOP JOIN | |11880 |15409|
|3 | EXCHANGE IN DISTR | |3 |37 |
|4 | EXCHANGE OUT DISTR (PKEY)|:EX10000|3 |37 |
|5 | PX PARTITION ITERATOR | |3 |37 |
|6 | TABLE SCAN |t4 |3 |37 |
|7 | PX PARTITION ITERATOR | |3960 |2561 |
|8 | TABLE SCAN |t2 |3960 |2561 |
===========================================================
Outputs & filters:
-------------------------------------
0 - output([t4.v1], [t4.v2], [t2.v1], [t2.v2]), filter(nil)
1 - output([t4.v1], [t4.v2], [t2.v1], [t2.v2]), filter(nil), dop=1
2 - output([t4.v1], [t4.v2], [t2.v1], [t2.v2]), filter(nil),
conds(nil), nl_params_([t4.v1])
3 - output([t4.v1], [t4.v2]), filter(nil)
4 - (#keys=1, [t4.v1]), output([t4.v1], [t4.v2]), filter(nil), dop=1
5 - output([t4.v1], [t4.v2]), filter(nil)
6 - output([t4.v1], [t4.v2]), filter(nil),
access([t4.v1], [t4.v2]), partitions(p[0-2])
7 - output([t2.v1], [t2.v2]), filter(nil)
8 - output([t2.v1], [t2.v2]), filter(nil),
access([t2.v1], [t2.v2]), partitions(p[0-3])
~~~
* 数据重分布
当联接键和左右表的分区键都没有关系的情况下，可以根据规则计算来选择使用 BROADCAST 还是 HASH HASH 的数据重分布方式，如下例所示：
**注意**
只有在并行度大于 1 时, 以下示例中两种数据重分发方式才有可能被选中。
~~~
obclient>EXPLAIN SELECT /*+ PARALLEL(2)*/* FROM t4, t2 WHERE t2.v2 = t4.v2\G;
*************************** 1. row ***************************
Query Plan:
=================================================================
|ID|OPERATOR |NAME |EST. ROWS|COST |
-----------------------------------------------------------------
|0 |PX COORDINATOR | |11880 |396863|
|1 | EXCHANGE OUT DISTR |:EX10001|11880 |394614|
|2 | HASH JOIN | |11880 |394614|
|3 | EXCHANGE IN DISTR | |3 |37 |
|4 | EXCHANGE OUT DISTR (BROADCAST)|:EX10000|3 |37 |
|5 | PX BLOCK ITERATOR | |3 |37 |
|6 | TABLE SCAN |t4 |3 |37 |
|7 | PX PARTITION ITERATOR | |400000 |256226|
|8 | TABLE SCAN |t2 |400000 |256226|
=================================================================
Outputs & filters:
-------------------------------------
0 - output([t4.v1], [t4.v2], [t2.v1], [t2.v2]), filter(nil)
1 - output([t4.v1], [t4.v2], [t2.v1], [t2.v2]), filter(nil), dop=2
2 - output([t4.v1], [t4.v2], [t2.v1], [t2.v2]), filter(nil),
equal_conds([t2.v2 = t4.v2]), other_conds(nil)
3 - output([t4.v1], [t4.v2]), filter(nil)
4 - output([t4.v1], [t4.v2]), filter(nil), dop=2
5 - output([t4.v1], [t4.v2]), filter(nil)
6 - output([t4.v1], [t4.v2]), filter(nil),
access([t4.v1], [t4.v2]), partitions(p[0-2])
7 - output([t2.v1], [t2.v2]), filter(nil)
8 - output([t2.v1], [t2.v2]), filter(nil),
access([t2.v1], [t2.v2]), partitions(p[0-3])
obclient>EXPLAIN SELECT /*+ PQ_DISTRIBUTE(t2 HASH HASH) PARALLEL(2)*/* FROM t4, t2
WHERE t2.v2 = t4.v2\G;
*************************** 1. row ***************************
Query Plan:
============================================================
|ID|OPERATOR |NAME |EST. ROWS|COST |
------------------------------------------------------------
|0 |PX COORDINATOR | |11880 |434727|
|1 | EXCHANGE OUT DISTR |:EX10002|11880 |432478|
|2 | HASH JOIN | |11880 |432478|
|3 | EXCHANGE IN DISTR | |3 |37 |
|4 | EXCHANGE OUT DISTR (HASH)|:EX10000|3 |37 |
|5 | PX BLOCK ITERATOR | |3 |37 |
|6 | TABLE SCAN |t4 |3 |37 |
|7 | EXCHANGE IN DISTR | |400000 |294090|
|8 | EXCHANGE OUT DISTR (HASH)|:EX10001|400000 |256226|
|9 | PX PARTITION ITERATOR | |400000 |256226|
|10| TABLE SCAN |t2 |400000 |256226|
============================================================
Outputs & filters:
-------------------------------------
0 - output([t4.v1], [t4.v2], [t2.v1], [t2.v2]), filter(nil)
1 - output([t4.v1], [t4.v2], [t2.v1], [t2.v2]), filter(nil), dop=2
2 - output([t4.v1], [t4.v2], [t2.v1], [t2.v2]), filter(nil),
equal_conds([t2.v2 = t4.v2]), other_conds(nil)
3 - output([t4.v1], [t4.v2]), filter(nil)
4 - (#keys=1, [t4.v2]), output([t4.v1], [t4.v2]), filter(nil), dop=2
5 - output([t4.v1], [t4.v2]), filter(nil)
6 - output([t4.v1], [t4.v2]), filter(nil),
access([t4.v1], [t4.v2]), partitions(p[0-2])
7 - output([t2.v1], [t2.v2]), filter(nil)
8 - (#keys=1, [t2.v2]), output([t2.v1], [t2.v2]), filter(nil), dop=2
9 - output([t2.v1], [t2.v2]), filter(nil)
10 - output([t2.v1], [t2.v2]), filter(nil),
access([t2.v1], [t2.v2]), partitions(p[0-3])
~~~
- 前言
- 1.说明
- 2.文档更新说明
- docker
- 01.docker安装
- 02.docker加速器
- 03.docker基本使用
- 04.docker 镜像与容器
- 05.Dockerfile
- 06.docker阿里镜像仓库
- 07.docker私有镜像仓库harbor
- 08.docker网络
- 09.docker项目实战01
- 10.docker项目实战02
- 11.docker componse
- 12.docker-compose常用命令
- 13.docker compose 案例
- 14.docker swarm集群
- 15.docker swarm常用命令
- 16.docker swarm 案例
- 17.volume
- 18.network
- 19.idea中部署项目到docker
- 20.docker目录方式挂载sqlite
- 21.docker常用命令补充
- 22.nginx容器代理静态文件403解决
- 23.docker集群管理平台
- k8s
- 01.Kubernetes介绍
- 02.K8s基本概念
- 03.K8s架构图
- 04.Minikube单节点环境搭建
- 05.kubeadm集群安装1.14.0
- 06.虚拟机静态网络配置
- 07.kubeadm高可用集群安装1.14.0
- 08 高可用VIP配置(keepalived+haproxy)
- 09.高可用免密登录
- 10.kubeadm init流程
- 11.k8s体验
- 12.网络插件
- 13.Ingress
- 14.Ingress分类
- 15.Dashboard
- 16.存储
- 01.Volumes
- 02.nfs
- 03.PV PVC
- 04.StorageClass
- 17.基础组件
- 01.Pod
- 02.Service
- 03.ReplicaSet(RS)
- 04.Deployment
- 06.Namespace
- 02.DaemonSet
- 03.StatefulSet
- 04.ReplicationController(RC)
- 06.Job
- 09.PetSet
- 10.StatefulSets
- 11.Federation
- 12.Secret
- 05.Resources
- 13.UserAccount/ServiceAccount
- 14.RBAC
- 18.核心组件
- Master组件
- 01.kube-apiserver
- 02.etcd
- 03.kube-controller-manager
- 04.cloud-controller-manager
- 05.kube-scheduler
- 06.DNS
- Node组件
- 01.kubelet
- 02.kube-proxy
- 03.docker
- 04.RKT
- 05.supervisord
- 06.fluentd
- kubectl
- 19.K8S服务更新部署
- 20.CI/CD
- 01.java安装
- 02.maven安装
- 03.gitlab安装
- 04.git安装
- 05.jenkins安装
- 06.k8s集群
- 07.DockerHub
- 08.实战
- 21.日志
- 01.不同组件日志
- 02.LogPilot+ES+Kibana
- 22.监控
- 23.k8s部署ocp项目[mysql]
- 01.ocp介绍
- 02.环境准备
- 03.镜像准备
- 04.部署说明
- 05.eureka-server
- 06.mysql
- 07.redis
- 08.auth-server
- 09.user-center
- 10.new-api-gateway
- 11.back-center
- 飞致云kubeoperator
- 01.kubeoperator介绍
- 02.kubeoperator安装
- 飞致云DataEase
- 项目介绍
- 系统架构
- 安装部署
- 在线安装
- 离线安装
- 用户手册
- 通用功能
- 数据源
- 数据集
- 视图
- 仪表板
- 系统管理
- 用户管理
- 飞致云JumpServer
- TIDB
- 网络
- 交换机
- ISO/OSI协议模型详解
- 交换CCNP
- RSTP快速生成树协议
- MST多生成树协议
- 以太网信道【应用广泛】
- 广播和多播抑制
- 多层交换
- ARP地址解析协议抑制
- VLAN间路由
- 热备份路由协议HSRP【思科私有】
- 虚拟路由器冗余协议VRRP
- linux
- 01.时间同步
- linux时间不能同步
- Linux挂载磁盘
- 安装ftp
- linux环境ftp账号
- HTTP状态码
- 宝塔
- Centos安装vsftp
- nginx ssl 配置
- datax
- 1.geom类型迁移扩展
- python安装
- 消息中间件
- 1.RocketMQ
- 1.RocketMQ单机环境安装
- 前端
- node踩坑之npm
- 数据库
- Mysql安装
- ClickHouse
- OceanBase数据库
- OceanBase介绍
- OceanBase数据库整体架构
- 快速入门
- 资源准备
- 安装 OBD部署 OceanBase 数据库
- 基本操作
- 数据库操作
- 表操作
- 索引操作
- 插入数据
- 删除数据
- 更新数据
- 提交事务
- 回滚事务
- 安装部署
- 使用 RPM 包安装 OceanBase 数据库
- 使用源码构建 OceanBase 数据库
- 设置无密码 SSH 登录
- 配置时钟源
- 数据分布
- 集群管理
- 租户与资源管理
- 数据分布1
- 数据副本与服务
- 数据均衡
- 数据模型
- 多租户架构
- 系统租户
- 普通租户
- 表格和表组
- 二级索引
- 无主键表
- 视图
- 高可用
- 高可用方案
- 部署模式
- redo 日志管理控制
- 事务管理
- 隔离级别
- 并发控制
- 全局时间戳服务
- 本地事务
- 分布式事务
- 分布式查询
- 存储架构
- LSM Tree 架构
- 内存表 MemTable
- 块存储 SSTable
- 转储和合并
- 缓存机制
- 读写流程
- DDL
- SQL 引擎
- SQL 请求执行流程
- 查询改写
- 基于规则的查询改写
- 基于代价的查询改写
- 查询优化
- 访问路径
- 基于规则的路径选择
- 基于代价的路径选择
- 联接算法
- 联接算法
- 联接顺序
- SQL 执行计划
- 执行计划算子
- TABLE SCAN
- TABLE LOOKUP
- JOIN
- COUNT
- GROUP BY
- WINDOW FUNCTION
- SUBPLAN FILTER
- DISTINCT
- SEQUENCE
- MATERIAL
- SORT
- LIMIT
- FOR UPDATE
- SELECT INTO
- SUBPLAN SCAN
- UNION
- INTERSECT
- EXCEPT/MINUS
- INSERT
- DELETE
- UPDATE
- MERGE
- EXCHANGE
- GI
- 执行计划缓存
- 快速参数化
- 实时执行计划展示
- 分布式执行计划
- 分布式执行和并行查询
- 分布式计划的生成
- 分布式执行计划调度
- 分布式执行计划管理
- 并行查询的执行
- 并行查询的参数调优
- 备份与恢复
- 备份架构
- 恢复架构
- Backup Set
- Archive Log Round
- 管理员指南
- 数据库基础组件介绍
- 数据库管理工具介绍
- OceanBase 客户端
- MySQL 客户端
- 数据库基础管理
- OceanBase 集群管理
- 集群参数管理
- 查询集群参数
- 修改集群参数
- Zone 管理
- 增加或删除 Zone
- 启动或停止 Zone
- 修改 Zone
- OBServer 管理
- 查看 OBServer 状态
- 停止 OBServer
- 启动 OBServer
- 管理 OBServer 节点状态
- 资源管理
- 创建资源单元
- 查看资源单元
- 修改资源单元
- 删除资源单元
- 创建资源池
- 查看资源配置
- 修改资源池
- 删除资源池
- 租户管理
- 创建用户租户
- 新建租户
- 查看租户
- 修改租户
- 删除租户
- 查看租户会话
- 终止租户会话
- 租户管理变量
- 内存管理
- OceanBase 内存结构
- OceanBase 数据库内存上限
- 系统内部内存管理
- 租户内部内存管理
- 执行计划缓存
- 常见内存问题
- 数据库对象管理
- 管理表
- 关于表
- 创建表
- 定义自增列
- 定义列的约束类型
- 查看表的定义
- 更改表
- 清空表
- 删除表
- 管理表组
- 关于表组
- 表组管理命令
- 管理索引
- 关于索引
- 创建索引
- 查看索引
- 删除索引
- 视图和同义词管理
- 管理视图
- 管理同义词
- 数据分布和链路管理
- 分区表和分区索引管理
- 关于分区
- 分区策略
- 创建分区表
- 一级分区表
- 二级分区表
- 维护分区表
- 一级分区表
- 二级分区表
- 分区裁剪
- 分区命名与查询
- 在分区表上建立索引
- 局部索引
- 全局索引
- 使用索引
- 副本管理
- 表级副本的使用
- Locality 管理
- 修改租户的 Locality
- 事务管理
- 提交事务
- 回滚事务
- 事务隔离级别
- 用户权限管理
- 创建用户
- 修改用户权限
- 查看白名单
- 锁定和解锁用户
- 删除用户
- 数据高可用
- 回收站管理
- 回收站支持的对象
- 数据库、表和索引级回收站
- 租户级回收站
- 物理备份与恢复管理
- 部署 NFS
- 备份数据
- 通过命令行备份
- 查看备份进度
- 停止备份
- 删除过期的备份
- 清理备份数据
- 取消清理备份数据
- 恢复数据
- 执行恢复
- 查看恢复进度和结果
- 备份维护
- 开发者指南
- 关于OceanBase数据库
- OceanBase 集群简介
- OceanBase 租户简介
- MySQL 租户数据库对象
- MySQL 客户端
- OceanBase 客户端（obclient）
- 关于结构化查询语言
- Java 数据库连接驱动（JDBC）
- 连接OceanBase数据库
- 通过 MySQL 客户端连接 OceanBase 租户
- 通过 obclient 连接 OceanBase 租户
- 创建 OceanBase 示例数据库 TPCC
- 通过 obclient 探索 OceanBase MySQL 租户
- 查询表数据
- 关于查询语句
- 查询表里符合特定搜索条件的数据
- 对查询的结果进行排序
- 从多个表里查询数据
- 在查询中使用操作符和函数
- 查看查询执行计划
- 在查询中使用 SQL Hint
- 关于查询超时设计
- 关于 DML 语句和事务
- 关于 DML 语句
- 关于 INSERT 语句
- 关于 UPDATE 语句
- 关于 DELETE 语句
- 关于 REPLACE INTO 语句
- 关于事务控制语句
- 提交事务
- 回滚事务
- 事务保存点
- 关于事务超时
- 创建和管理数据库对象
- 关于 DDL 语句
- 创建数据库
- 创建和管理表
- 关于 SQL 数据类型
- 创建表
- 关于自增列
- 关于列的约束类型
- 关于表的索引
- 闪回被删除的表
- 创建和管理分区表
- 分区路由
- 分区策略
- 分区表的索引
- 分区表使用建议
- 创建和管理表组
- 关于表组
- 创建表时指定表组
- 查看表组信息
- 向表组中增加表
- 删除表组
- 创建和管理视图
- 创建视图
- 修改视图
- 删除视图
- 向 OceanBase 迁移数据
- DataX
- 不同数据源的 DataX 读写插件示例
- OceanBase 数据加载技术
- 附录
- OceanBase 常用参数和变量
- OceanBase 常用 SQL
- SQL参考
- 基本元素
- 运算符
- 函数
- 函数
- 聚集函数
- 分析函数
- 信息函数
- 其它函数
- 查询和子查询
- 连接
- 集合
- SQL语句
- 通用语法
- ALTER DATABASE
- ALTER OUTLINE
- ALTER RESOURCE POOL
- ALTER RESOURCE UNIT
- ALTER SYSTEM
- ALTER TABLE
- ALTER TABLEGROUP
- ALTER TENANT
- ALTER USER
- CREATE DATABASE
- CREATE INDEX
- CREATE OUTLINE
- CREATE RESOURCE POOL
- CREATE RESOURCE UNIT
- CREATE RESTORE POINT
- CREATE SYNONYM
- CREATE TABLE
- CREATE TABLEGROUP
- CREATE TENANT
- CREATE USER
- CREATE VIEW
- DELETE
- DROP DATABASE
- DROP INDEX
- DROP OUTLINE
- DROP RESOURCE POOL
- DROP RESOURCE UNIT
- DROP RESTORE POINT
- DROP TABLE
- DROP TABLEGROUP
- DROP TENANT
- DROP SYNONYM
- DROP USER
- DROP VIEW
- EXPLAIN
- FLASHBACK DATABASE
- FLASHBACK TABLE
- FLASHBACK TENANT
- GRANT
- INSERT
- KILL
- PURGE DATABASE
- PURGE INDEX
- PURGE RECYCLEBIN
- PURGE TABLE
- PURGE TENANT
- RENAME TABLE
- RENAME USER
- REPLACE
- REVOKE
- SAVEPOINT
- SCHEMA
- SELECT
- SESSION
- SET PASSWORD
- SHOW GRANTS
- SHOW RECYCLEBIN
- TRANSACTION
- TRUNCATE TABLE
- UPDATE
- SQL调优指南
- SQL请求执行流程
- SQL 执行计划
- SQL 执行计划简介
- 执行计划算子
- TABLE SCAN
- TABLE LOOKUP
- JOIN
- COUNT
- GROUP BY
- WINDOW FUNCTION
- SUBPLAN FILTER
- DISTINCT
- SEQUENCE
- MATERIAL
- SORT
- LIMIT
- FOR UPDATE
- SELECT INTO
- SUBPLAN SCAN
- UNION
- INTERSECT
- EXCEPT/MINUS
- INSERT
- DELETE
- UPDATE
- MERGE
- EXCHANGE
- GI
- 执行计划缓存
- 快速参数化
- 实时执行计划展示
- 分布式执行计划
- 分布式执行和并行查询
- 分布式计划的生成
- 分布式执行计划调度
- 分布式执行计划管理
- 并行查询的执行
- 并行查询的参数调优
- 参考指南
- 系统视图
- 概述
- 字典视图
- mysql.help_topic
- mysql.help_category
- mysql.help_keyword
- mysql.help_relation
- mysql.db
- mysql.proc
- mysql.time_zone
- mysql.time_zone_name
- mysql.time_zone_transition
- mysql.time_zone_transition_type
- mysql.user
- information_schema.CHARACTER_SETS
- information_schema.COLLATIONS
- information_schema.COLLATION_CHARACTER_SET_APPLICABILITY
- information_schema.COLUMNS
- information_schema.DBA_OUTLINES
- information_schema.ENGINES
- information_schema.GLOBAL_STATUS
- information_schema.GLOBAL_VARIABLES
- information_schema.KEY_COLUMN_USAGE
- information_schema.PARAMETERS
- information_schema.PARTITIONS
- information_schema.PROCESSLIST
- information_schema.REFERENTIAL_CONSTRAINTS
- information_schema.ROUTINES
- information_schema.SCHEMATA
- information_schema.SCHEMA_PRIVILEGES
- information_schema.SESSION_STATUS
- information_schema.SESSION_VARIABLES
- information_schema.STATISTICS
- information_schema.TABLES
- information_schema.TABLE_CONSTRAINTS
- information_schema.TABLE_PRIVILEGES
- information_schema.USER_PRIVILEGES
- information_schema.USER_RECYCLEBIN
- information_schema.VIEWS
- oceanbase.CDB_OB_BACKUP_ARCHIVELOG_SUMMARY
- oceanbase.CDB_OB_BACKUP_JOB_DETAILS
- oceanbase.CDB_OB_BACKUP_SET_DETAILS
- oceanbase.CDB_OB_BACKUP_PROGRESS
- oceanbase.CDB_OB_BACKUP_SET_EXPIRED
- oceanbase.CDB_OB_BACKUP_ARCHIVELOG_PROGRESS
- oceanbase.CDB_OB_BACKUP_CLEAN_HISTORY
- oceanbase.CDB_OB_BACKUP_TASK_CLEAN_HISTORY
- oceanbase.CDB_OB_RESTORE_PROGRESS
- oceanbase.CDB_OB_RESTORE_HISTORY
- oceanbase.CDB_CKPT_HISTORY
- oceanbase.CDB_OB_BACKUP_VALIDATION_JOB
- oceanbase.CDB_OB_BACKUP_VALIDATION_JOB_HISTORY
- oceanbase.CDB_OB_TENANT_BACKUP_VALIDATION_TASK
- oceanbase.CDB_OB_BACKUP_VALIDATION_TASK_HISTORY
- oceanbase.CDB_OB_BACKUP_BACKUP_ARCHIVELOG_SUMMARY
- oceanbase.CDB_OB_BACKUP_BACKUPSET_TASK_HISTORY
- oceanbase.CDB_OB_BACKUP_BACKUPSET_TASK
- oceanbase.CDB_OB_BACKUP_BACKUPSET_JOB_HISTORY
- oceanbase.CDB_OB_BACKUP_BACKUPSET_JOB
- oceanbase.CDB_OB_BACKUP_SET_OBSOLETE
- 性能视图
- gv$plan_cache_stat
- gv$plan_cache_plan_stat
- gv$session_event
- gv$session_wait
- gv$session_wait_history
- gv$system_event
- gv$sesstat
- gv$sysstat
- gv$sql_audit
- gv$latch
- gv$memory
- gv$memstore
- gv$memstore_info
- gv$plan_cache_plan_explain
- gv$obrpc_outgoing
- gv$obrpc_incoming
- gv$sql
- gv$sql_plan_monitor
- gv$outline
- gv$concurrent_limit_sql
- gv$sql_plan_statistics
- gv$server_memstore
- gv$unit_load_balance_event_history
- gv$tenant
- gv$database
- gv$table
- gv$unit
- gv$partition
- gv$lock_wait_stat
- gv$session_longops
- gv$tenant_memstore_allocator_info
- gv$minor_merge_info
- gv$tenant_px_worker_stat
- gv$partition_audit
- gv$ps_stat
- gv$ps_item_info
- gv$sql_workarea
- gv$sql_workarea_histogram
- gv$ob_sql_workarea_memory_info
- gv$server_schema_info
- gv$merge_info
- gv$lock
- gv$sstable
- gv$ob_trans_table_status
- v$statname
- v$event_name
- v$session_event
- v$session_wait
- v$session_wait_history
- v$sesstat
- v$sysstat
- v$system_event
- v$memory
- v$memstore
- v$memstore_info
- v$plan_cache_stat
- v$plan_cache_plan_stat
- v$plan_cache_plan_explain
- v$sql_audit
- v$obrpc_outgoing
- v$obrpc_incoming
- v$sql
- v$sql_monitor
- v$sql_plan_monitor
- v$sql_plan_statistics
- v$unit
- v$partition
- v$lock_wait_stat
- v$session_longops
- v$latch
- v$tenant_memstore_allocator_info
- v$tenant_px_worker_stat
- v$partition_audit
- v$ob_cluster
- v$ob_standby_status
- v$ob_cluster_stats
- v$ob_cluster_event_history
- v$ps_stat
- v$ps_item_info
- v$sql_workarea
- v$sql_workarea_active
- v$sql_workarea_histogram
- v$ob_sql_workarea_memory_info
- v$ob_timestamp_service
- v$server_schema_info
- v$merge_info
- v$lock
- v$sql_monitor_statname
- v$restore_point
- v$ob_cluster_failover_info
- v$encrypted_tables
- v$encrypted_tablespaces
- v$sstable
- v$ob_trans_table_status
- 系统变量
- 系统变量概述
- auto_increment_increment
- auto_increment_offset
- autocommit
- character_set_client
- character_set_connection
- character_set_database
- character_set_results
- character_set_server
- character_set_system
- collation_connection
- collation_database
- collation_server
- interactive_timeout
- last_insert_id
- max_allowed_packet
- sql_mode
- time_zone
- tx_isolation
- version_comment
- wait_timeout
- binlog_row_image
- character_set_filesystem
- connect_timeout
- datadir
- debug_sync
- div_precision_increment
- explicit_defaults_for_timestamp
- group_concat_max_len
- identity
- lower_case_table_names
- net_read_timeout
- net_write_timeout
- read_only
- sql_auto_is_null
- sql_select_limit
- timestamp
- tx_read_only
- version
- sql_warnings
- max_user_connections
- init_connect
- license
- net_buffer_length
- system_time_zone
- query_cache_size
- query_cache_type
- sql_quote_show_create
- max_sp_recursion_depth
- sql_safe_updates
- ob_proxy_partition_hit
- ob_log_level
- ob_max_parallel_degree
- ob_query_timeout
- ob_read_consistency
- ob_enable_transformation
- ob_trx_timeout
- ob_enable_plan_cache
- ob_enable_index_direct_select
- ob_proxy_set_trx_executed
- ob_enable_aggregation_pushdown
- ob_last_schema_version
- ob_global_debug_sync
- ob_proxy_global_variables_version
- ob_enable_trace_log
- ob_enable_hash_group_by
- ob_enable_blk_nestedloop_join
- ob_bnl_join_cache_size
- ob_org_cluster_id
- ob_plan_cache_percentage
- ob_plan_cache_evict_high_percentage
- ob_plan_cache_evict_low_percentage
- recyclebin
- ob_capability_flag
- ob_stmt_parallel_degree
- is_result_accurate
- error_on_overlap_time
- ob_compatibility_mode
- ob_create_table_strict_mode
- ob_sql_work_area_percentage
- ob_route_policy
- ob_enable_transmission_checksum
- foreign_key_checks
- ob_enable_truncate_flashback
- ob_tcp_invited_nodes
- sql_throttle_current_priority
- sql_throttle_priority
- sql_throttle_rt
- sql_throttle_network
- auto_increment_cache_size
- ob_enable_jit
- ob_timestamp_service
- plugin_dir
- undo_retention
- ob_sql_audit_percentage
- ob_enable_sql_audit
- optimizer_use_sql_plan_baselines
- optimizer_capture_sql_plan_baselines
- parallel_max_servers
- parallel_servers_target
- ob_trx_idle_timeout
- block_encryption_mode
- ob_reserved_meta_memory_percentage
- ob_check_sys_variable
- tracefile_identifier
- transaction_isolation
- ob_trx_lock_timeout
- validate_password_check_user_name
- validate_password_length
- validate_password_mixed_case_count
- validate_password_number_count
- validate_password_policy
- validate_password_special_char_count
- default_password_lifetime
- ob_trace_info
- secure_file_priv
- ob_pl_block_timeout
- performance_schema
- transaction_read_only
- resource_manager_plan
- 系统配置项
- 系统配置项概述
- auto_leader_switch_interval
- auto_delete_expired_backup
- autoinc_cache_refresh_interval
- audit_sys_operations
- audit_trail
- balancer_idle_time
- balancer_log_interval
- balancer_timeout_check_interval
- balancer_task_timeout
- balancer_tolerance_percentage
- balancer_emergency_percentage
- balance_blacklist_failure_threshold
- balance_blacklist_retry_interval
- backup_concurrency
- backup_dest
- backup_net_limit
- backup_recovery_window
- backup_region
- builtin_db_data_verify_cycle
- bf_cache_miss_count_threshold
- bf_cache_priority
- cache_wash_threshold
- clog_cache_priority
- clog_sync_time_warn_threshold
- clog_disk_usage_limit_percentage
- clog_transport_compress_all
- clog_transport_compress_func
- clog_persistence_compress_func
- clog_max_unconfirmed_log_count
- cluster
- cluster_id
- cpu_count
- cpu_quota_concurrency
- cpu_reserved
- config_additional_dir
- data_copy_concurrency
- data_dir
- datafile_disk_percentage
- dtl_buffer_size
- datafile_size
- debug_sync_timeout
- default_compress_func
- default_compress
- default_progressive_merge_num
- default_row_format
- devname
- data_disk_usage_limit_percentage
- disk_io_thread_count
- dead_socket_detection_timeout
- enable_clog_persistence_compress
- election_cpu_quota
- enable_one_phase_commit
- enable_sys_unit_standalone
- enable_pg
- enable_smooth_leader_switch
- election_blacklist_interval
- enable_election_group
- enable_auto_leader_switch
- enable_global_freeze_trigger
- enable_manual_merge
- enable_merge_by_turn
- enable_perf_event
- enable_rebalance
- enable_record_trace_log
- enable_record_trace_id
- enable_early_lock_release
- enable_rereplication
- enable_rich_error_msg
- enable_rootservice_standalone
- enable_sql_audit
- enable_sql_operator_dump
- enable_async_syslog
- enable_syslog_recycle
- enable_syslog_wf
- enable_upgrade_mode
- enable_separate_sys_clog
- enable_ddl
- enable_major_freeze
- enable_rebuild_on_purpose
- enable_log_archive
- enable_monotonic_weak_read
- external_kms_info
- freeze_trigger_percentage
- flush_log_at_trx_commit
- fuse_row_cache_priority
- force_refresh_location_cache_interval
- force_refresh_location_cache_threshold
- get_leader_candidate_rpc_timeout
- global_major_freeze_residual_memory
- global_write_halt_residual_memory
- ignore_replay_checksum_error
- global_index_build_single_replica_timeout
- high_priority_net_thread_count
- ignore_replica_checksum_error
- ignore_replay_checksum_error
- index_cache_priority
- index_clog_cache_priority
- index_info_block_cache_priority
- internal_sql_execute_timeout
- large_query_worker_percentage
- large_query_threshold
- leak_mod_to_check
- lease_time
- location_cache_cpu_quota
- location_cache_expire_time
- location_cache_priority
- location_cache_refresh_min_interval
- location_fetch_concurrency
- location_refresh_thread_count
- log_archive_checkpoint_interval
- log_archive_concurrency
- log_restore_concurrency
- major_freeze_duty_time
- max_kept_major_version_number
- max_string_print_length
- max_syslog_file_count
- merge_stat_sampling_ratio
- major_compact_trigger
- memory_chunk_cache_size
- memory_limit
- memory_limit_percentage
- memory_reserved
- merge_thread_count
- merger_check_interval
- merger_completion_percentage
- merger_switch_leader_duration_time
- merger_warm_up_duration_time
- max_px_worker_count
- migration_disable_time
- min_observer_version
- minor_deferred_gc_time
- minor_freeze_times
- minor_warm_up_duration_time
- mysql_port
- minor_merge_concurrency
- multiblock_read_gap_size
- multiblock_read_size
- micro_block_merge_verify_level
- migrate_concurrency
- minor_compact_trigger
- memstore_limit_percentage
- net_thread_count
- obconfig_url
- ob_enable_batched_multi_statement
- partition_table_check_interval
- partition_table_scan_batch_count
- plan_cache_evict_interval
- px_task_size
- px_workers_per_cpu_quota
- replica_safe_remove_time
- resource_hard_limit
- resource_soft_limit
- rootservice_async_task_queue_size
- rootservice_async_task_thread_count
- rootservice_list
- rootservice_ready_check_interval
- row_compaction_update_limit
- row_purge_thread_count
- rpc_port
- rpc_timeout
- restore_concurrency
- rootservice_memory_limit
- rebuild_replica_data_lag_threshold
- schema_history_expire_time
- ssl_client_authentication
- server_check_interval
- server_data_copy_in_concurrency
- server_data_copy_out_concurrency
- server_permanent_offline_time
- stack_size
- server_balance_critical_disk_waterlevel
- server_balance_disk_tolerance_percent
- system_memory
- server_balance_cpu_mem_tolerance_percent
- server_cpu_quota_max
- server_cpu_quota_min
- sql_audit_memory_limit
- sys_bkgd_io_high_percentage
- sys_bkgd_io_low_percentage
- sys_bkgd_io_timeout
- sys_bkgd_net_percentage
- sys_bkgd_migration_change_member_list_timeout
- sys_bkgd_migration_retry_num
- syslog_level
- switchover_process_thread_count
- system_cpu_quota
- sys_cpu_limit_trigger
- syslog_io_bandwidth_limit
- tablet_size
- tableapi_transport_compress_func
- tenant_task_queue_size
- tenant_groups
- trace_log_slow_query_watermark
- trace_log_sampling_interval
- trx_2pc_retry_interval
- trx_force_kill_threshold
- tde_method
- token_reserved_percentage
- unit_balance_resource_weight
- user_block_cache_priority
- user_row_cache_priority
- user_tab_col_stat_cache_priority
- user_iort_up_percentage
- use_large_pages
- virtual_table_location_cache_expire_time
- workers_per_cpu_quota
- wait_leader_batch_count
- writing_throttling_maximum_duration
- writing_throttling_trigger_percentage
- weak_read_version_refresh_interval
- workarea_size_policy
- zone
- zone_merge_concurrency
- zone_merge_order
- zone_merge_timeout
- ob_ssl_invited_common_names
- ssl_external_kms_info
- ob_event_history_recycle_interval
- backup_log_archive_checkpoint_interval
- plsql_ccflags
- plsql_code_type
- plsql_debug
- plsql_optimize_level
- plsql_v2_compatibility
- plsql_warnings
- recyclebin_object_expire_time
- log_archive_batch_buffer_limit
- clog_disk_utilization_threshold
- backup_backup_archive_log_batch_count
- backup_backup_archivelog_retry_interval
- backup_backupset_batch_count
- backup_backupset_retry_interval
- open_cursors
- fast_recovery_concurrency
- 预留关键字
- 部署实践
- 设置无密码 SSH 登录
- 单机安装
- 本地安装
- 分布式安装
- 创建租户
- OceanBaseDeploy(OBD)常用命令
- 大数据
- 数据仓库分层
- 数据仓库分层实践
- hive安装
- hive命令
- hadoop安装
- jdk安装
- 应龙inlong
- 网关
- apisix
- apisix2.7源码安装
- apisix rpm2.6安装
- apisix-dashboard2.7 rpm安装
- apisix-dashboard使用
- apisix-dashboard进阶