PanWeiDB2.0异构数据库访问测试

PanWeiDB2.0异构数据库访问测试

异构数据库访问兼容性测试一览表

No	访问路径	多维度结果	备注
1	PanWeiDB（集中式）—访问—PanWeiDB（集中式）	√	支持复杂SQL
2	PanWeiDB（集中式）—访问—Oracle	√	支持
3	Oracle—访问—PanWeiDB（集中式）	√	受Oracle默认大写影响，dblink涉及磐维的对象及字段，需要用双引号""括起来。不支持求平均值，偶发性出现链接错误，重新登录正常
4	PanWeiDB（集中式）—访问—PanWeiDB（分布式）	×	基础select不支持，缺少get_typmod_with_unit系统基础函数
5	PanWeiDB（分布式）—访问—PanWeiDB（集中式）	×	不支持database link语法，厂商反馈：不支持database link访问磐维集中式
6	PanWeiDB（集中式）—访问—GoldenDB（分布式）	√	支持简单SQL，需注意与GoldenDB数据交互或者修改时，外部表定义的数据类型可能会出现转换异常的情况
7	PanWeiDB（集中式）—访问—AntDB（集中式）	×	认证交互错误：DETAIL: expected authentication request from server, but received v
8	AntDB（集中式）—访问—PanWeiDB（集中式）	×	外部表映射错误，ERROR: user mapping not found for “antdb”

测试过程

1、PanWeiDB访问PanWeiDB

测试结果

序号	复杂度	具体操作	结果
1	基本SQL操作	查询远程数据（SELECT）	验证通过
2		插入数据到本地表（INSERT）	验证通过
3		更新本地表数据（UPDATE）	验证通过
4		删除数据（DELETE）	验证通过
5	复杂SQL操作	多表关联（4表JOIN）	验证通过
6		笛卡尔积与过滤条件	验证通过
7		嵌套循环（NESTED LOOP）	验证通过
8		哈希连接（HASH JOIN）	验证通过
9		聚合函数与GROUP BY	验证通过
10		子查询与EXISTS	验证通过
11		窗口函数与RANK	验证通过
12		复杂条件与函数	验证通过

测试账户

create user testuser1 with sysadmin password '3GahEH271';
create database testdb1 owner testuser1;
\q
gsql -r -p 17700 -d testdb1 -U testuser1 -W 3GahEH271

1.1 创建dblink

grant all on database testdb1 to testuser1;
GRANT ALL PRIVILEGES ON DATABASE testdb1 TO testuser1;

create extension postgres_fdw;
grant usage on foreign data wrapper postgres_fdw to testuser1;

# 创建到oracle的映射，执行此语句需预先使用 gs_guc generate 命令生成 datasource 文件，此处以如下命令为例：
gs_guc generate -S 'Gs@123456' -D $GAUSSHOME/bin -o usermapping

create database link dblink_pwcrm149 connect to testuser1 identified by '3GahEH271' using postgres_fdw(host '10.183.162.149',port '17700',dbname 'testdb1');

2.1 基本SQL操作

3. 复杂SQL示例

3.1 多表关联（4表JOIN）

3.2 笛卡尔积与过滤条件

3.3 嵌套循环（NESTED LOOP）

3.4 哈希连接（HASH JOIN）

3.5 聚合函数与GROUP BY

3.6 子查询与EXISTS

3.7 窗口函数与RANK

3.8 复杂条件与函数

2、PanWeiDB访问Oracle

PanWeiDB（PanWeiDB版本升级测试库），访问，Oracle（运维库）

create extension oracle_fdw;
create user cboss_ora password 'cboss_ora@123';
grant usage on foreign data wrapper oracle_fdw to cboss_ora ;
\c - cboss_ora
create server ora_fdw_server foreign data wrapper oracle_fdw
options(dbserver '10.176.240.129:1521/cbosslo');
# gs_guc generate -S 'cboss_ora@123' -D $GAUSSHOME/bin -o usermapping
create user mapping for cboss_ora server ora_fdw_server options(user 'OPERUSER_MY',password 'zvMc!P47dK');
create foreign table emp_fdw_ora(empno int,ename varchar(30)) server ora_fdw_server options(schema 'SYSTEM',table 'EMP_FDW');
create foreign table emp_fdw_ora(empno int,ename varchar(30)) server ora_fdw_server options(schema 'OPERUSER_MY',table 'emp_fdw');
select * from emp_fdw_ora;
insert into emp_fdw_ora values(3,'bar3');

测试复杂场景

drop table t1;
create table t1 (id number primary key,update_time timestamp);
INSERT INTO t1 (id, update_time) VALUES (1, TO_TIMESTAMP('2025-01-01 10:00:00', 'YYYY-MM-DD HH24:MI:SS'));
INSERT INTO t1 (id, update_time) VALUES (2, TO_TIMESTAMP('2025-01-02 12:00:00', 'YYYY-MM-DD HH24:MI:SS'));
INSERT INTO t1 (id, update_time) VALUES (3, TO_TIMESTAMP('2025-01-03 14:00:00', 'YYYY-MM-DD HH24:MI:SS'));
commit;
select * from t1;
drop table t2;
create table t2 (
    id number,
    name varchar2(50),
    t1_id number
);
INSERT INTO t2 (id, name, t1_id) VALUES (101, 'Alice', 1);
INSERT INTO t2 (id, name, t1_id) VALUES (102, 'Bob', 2);
INSERT INTO t2 (id, name, t1_id) VALUES (103, 'Charlie', 3);
commit;
select * from t2;
select * from t1_fdw_ora;
SELECT t1.id, t1.update_time, t2.name
FROM t1_fdw_ora as t1
JOIN t2 as t2 ON t1.id = t2.t1_id;
testdb1=> explain analyze SELECT t1.id, t1.update_time, t2.name
testdb1-> FROM t1_fdw_ora as t1
testdb1-> JOIN t2 as t2 ON t1.id = t2.t1_id;
                                                         QUERY PLAN
----------------------------------------------------------------------------------------------------------------------------
 Hash Join  (cost=100.04..114.88 rows=2 width=158) (actual time=1.364..1.404 rows=3 loops=1)
   Hash Cond: (t2.t1_id = t1.id)
   ->  Seq Scan on t2  (cost=0.00..13.85 rows=385 width=150) (actual time=0.023..0.034 rows=3 loops=1)
   ->  Hash  (cost=100.03..100.03 rows=1 width=40) (actual time=1.215..1.215 rows=3 loops=1)
          Buckets: 32768  Batches: 1  Memory Usage: 257kB
         ->  Foreign Scan on t1_fdw_ora t1  (cost=100.00..100.03 rows=1 width=40) (actual time=0.904..1.165 rows=3 loops=1)
 Total runtime: 20.764 ms
(7 rows)

testdb1=>

3、Oracle访问PanWeiDB

Oracle（DBA运维库），访问，PanWeiDB（PanWeiDB版本升级测试库）

Oracle使用透明网关访问PanWeiDB数据库

3.1 多表关联（4表JOIN）

-- 左连接 + 右连接 + 窗口函数
select
    "ee"."emp_name",
    d.dept_name,
    "pp"."project_name",
    pa.role,
    "ee"."salary"
from "employees"@pg_link "ee"
left join departments d on "ee"."department" = d.dept_name
right join "projects"@pg_link "pp" on d.dept_id = "pp"."dept_id"
join project_assignments pa on pa.project_id = "pp"."project_id"
where "pp"."start_date" > sysdate - 365
and "ee"."salary" > 6000;

database link涉及磐维对象及字段，需要用双引号""括起来，这受Oracle默认大写影响

不支持求平均值

可以查询

偶发性出现链接错误

3.2 笛卡尔积与过滤条件

-- 笛卡尔积 + CASE判断
SELECT
    "ee"."emp_name",
    p.project_name,
    CASE
        WHEN "ee"."salary" > 10000 THEN 'High'
        WHEN "ee"."salary" BETWEEN 5000 AND 10000 THEN 'Medium'
        ELSE 'Low'
    END AS salary_level
FROM "employees"@pg_link "ee", projects p
WHERE p.dept_id = 1
AND "ee"."department" = 'IT';

执行计划

3.3 嵌套循环（NESTED LOOP）

-- 强制使用NESTED LOOP（提示方式）
SELECT /*+ USE_NL(e "dd") */
    e.emp_id,
    "dd"."dept_name"
FROM employees e
JOIN "departments"@pg_link "dd" ON e.department = "dd"."dept_name"
WHERE "dd"."budget" > 500000;

3.4 哈希连接（HASH JOIN）

-- 强制使用HASH JOIN
SELECT /*+ USE_HASH(pa "pp") */
    pa.assignment_id,
    "pp"."project_name"
FROM project_assignments pa
JOIN "projects"@pg_link "pp" ON pa.project_id = "pp"."project_id"
WHERE "pp"."end_date" > SYSDATE;

3.5 聚合函数与GROUP BY

-- 聚合函数 + HAVING
SELECT
    "dd"."dept_name",
    COUNT(e.emp_id) AS total_employees,
    SUM(e.salary) AS total_salary
FROM "departments"@pg_link "dd"
LEFT JOIN employees e ON "dd"."dept_name" = e.department
GROUP BY "dd"."dept_name"
HAVING SUM(e.salary) > 100000;

3.6 子查询与EXISTS

-- EXISTS子查询
SELECT
    e.emp_name
FROM employees e
WHERE EXISTS (
    SELECT 1
    FROM "project_assignments"@pg_link "pa"
    WHERE "pa"."emp_id" = e.emp_id
    AND "pa"."role" = 'Manager'
);

3.7 窗口函数与RANK

-- RANK()窗口函数
SELECT
    "emp_name",
    "salary",
    RANK() OVER (ORDER BY "salary" DESC) AS salary_rank
FROM "employees"@pg_link
WHERE "department" = 'IT';

3.8 复杂条件与函数

-- 字符串函数 + 日期计算
SELECT
    "ee"."emp_name",
    UPPER(d.dept_name) AS dept_upper,
    ADD_MONTHS("ee"."hire_date", 12) AS hire_anniversary
FROM "employees"@pg_link "ee"
JOIN departments d ON "ee"."department" = d.dept_name
WHERE TO_CHAR("ee"."hire_date", 'YYYY') = '2023'
AND LENGTH("ee"."emp_name") > 5;

4、集中式PanWeiDB访问分布式PanWeiDB

PanWeiDB（PanWeiDB版本升级测试库），访问，PanWeiDB分布式（计费系统）

测试结果

序号	复杂度	具体操作	结果
1	基本SQL操作	查询远程数据（SELECT）	测试未通过

测试账户

create user testuser2 with sysadmin password '3GahEH271';
create database testdb1 owner testuser2;
\q
gsql -r -p 17700 -d testdb1 -U testuser2 -W 3GahEH271

1.1 创建dblink

create database link dblink_jfjsxt19098 connect to testuser2 identified by '3GahEH271' using postgres_fdw(host '10.183.190.98',port '17700',dbname 'testdb1');

2.1 基本SQL操作

testdb1=> SELECT * FROM employees@dblink_jfjsxt19098 WHERE department = 'IT';
ERROR:  function get_typmod_with_unit(integer, text[]) does not exist
HINT:  No function matches the given name and argument types. You might need to add explicit type casts.
CONTEXT:  Remote SQL command: SELECT a.attname,
  (case pg_catalog.format_type(a.atttypid, get_typmod_with_unit(a.atttypmod, a.attoptions)) when 'oradate' then 'date' else pg_catalog.format_type(a.atttypid, get_typmod_with_unit(a.atttypmod, a.attoptions)) end)
  FROM pg_catalog.pg_attribute a
  WHERE a.attrelid = (select oid from pg_class where relname='employees' and relnamespace=(select oid from pg_namespace where nspname='testuser2'))
  AND a.attnum > 0 AND NOT a.attisdropped AND a.attkvtype != 4 AND a.attname <> 'tableoid' AND a.attname <> 'tablebucketid'
  ORDER BY a.attnum;
referenced column: format_type
testdb1=>

磐维集中式数据库-SQL执行正常

磐维分布式数据库-SQL无法执行

磐维分布式数据库缺少函数“get_typmod_with_unit”，导致集中式数据库访问分布式数据库时，无法执行查询：

查询 pg_attribute 系统表，检索有关表 employees 的列信息，以获取列名和数据类型信息，

get_typmod_with_unit 是一个系统函数，用于从类型修饰符中提取单位信息。

它的作用是将类型修饰符转换为更易读的格式

手动创建函数get_typmod_with_unit

CREATE OR REPLACE FUNCTION get_typmod_with_unit(typmod integer, options text[])
RETURNS integer
AS $$
DECLARE
    unit integer;
BEGIN
    -- 从类型修饰符中提取单位信息
    -- 这里是一个示例逻辑，实际逻辑可能需要根据你的数据库系统进行调整
    unit := typmod % 1000; -- 假设单位信息存储在 typmod 的低三位

    -- 返回提取的单位信息
    RETURN unit;
END;
$$ LANGUAGE plpgsql;

磐维集中式dblink访问

查询报错，依然缺少函数“get_typmod_with_unit”函数，暂无方案解决（20250305）。

5、分布式PanWeiDB访问集中式PanWeiDB

PanWeiDB分布式（结算系统（计费系统-计费云详单）），访问，PanWeiDB（PanWeiDB版本升级测试库）

分布式磐维数据库，暂不支持dblink

6、PanWeiDB访问GoldenDB

PanWeiDB（PanWeiDB版本升级测试库），访问，GoldenDB分布式（中台中心）

数据类型略微调整

create extension mysql_fdw;
grant usage on foreign data wrapper mysql_fdw to testuser1;
create server goldendb_fdw_server foreign data wrapper mysql_fdw options(HOST '10.183.200.193',PORT '8880');
create user mapping for testuser1 server goldendb_fdw_server options(username 'testuser1',password '3GahEH271!');
create foreign table employees_fdw_goldendb (
    emp_id NUMBER(10) ,
    emp_name VARCHAR2(50) NOT NULL,
    hire_date DATE,
    salary NUMBER(10, 2),
    department VARCHAR2(50),
    resume CLOB,
    photo BLOB
) server goldendb_fdw_server options(DBNAME 'testdb1',table_name 'employees');
create foreign table departments_fdw_goldendb (
    dept_id NUMBER(10),
    dept_name VARCHAR2(50) ,
    manager_id NUMBER(10),
    budget NUMBER(15, 2)
) server goldendb_fdw_server options(DBNAME 'testdb1',table_name 'departments');
create foreign table projects_fdw_goldendb (
    project_id NUMBER(10),
    project_name VARCHAR2(100),
    start_date DATE,
    end_date DATE,
    dept_id NUMBER(10)
) server goldendb_fdw_server options(DBNAME 'testdb1',table_name 'projects');
create foreign table project_assignments_fdw_goldendb (
    assignment_id NUMBER(10) ,
    emp_id NUMBER(10),
    project_id NUMBER(10),
    role VARCHAR2(50)
) server goldendb_fdw_server options(DBNAME 'testdb1',table_name 'project_assignments');

基本SQL操作

根据错误信息，问题出在数据类型转换上。具体来说，磐维无法将常量值的数据类型（1399）转换为GoldenDB 的数据类型。这通常发生在使用外部表（FDW，Foreign Data Wrapper）时，数据类型不匹配。

错误分析

1. 数据类型不匹配

   • 错误信息 cannot convert constant value to MySQL value 表明，磐维 无法将某个常量值转换为 GoldenDB 的数据类型。
   • Constant value data type: 1399 表示常量值的数据类型是 1399，这可能是 GoldenDB 中的 DECIMAL 类型。

2. 外部表定义问题

   • 外部表 employees_fdw_goldendb 的定义可能与GoldenDB 表的结构不完全匹配。

解决方案

1. 检查外部表定义

   • 确保外部表 employees_fdw_goldendb 的定义与GoldenDB 表的结构完全匹配。
   • 使用 SHOW CREATE TABLE 或 DESCRIBE 命令检查 GoldenDB 表的结构。

2. 修改外部表定义

   • 如果数据类型不匹配，修改外部表的定义，使其与 GoldenDB 表的结构一致。
   • 例如，如果 GoldenDB 表中的 emp_id 是 INT 类型，确保外部表中的 emp_id 也是 INT 类型。

3. 检查常量值的数据类型

   • 确保插入或查询时使用的常量值的数据类型与外部表的定义匹配。
   • 如果需要，使用显式类型转换。

MySQL_FDW功能描述

外部表，支持select查询、dml操作。

支持创建外部数据封装器mysql_fdw，连接MariaDB或MySQL或者GoldenDB数据库，并能在外部表上进行查询、插入、更新和删除操作。

mysql_fdw插件默认参与编译，用户可直接使用mysql_fdw，无须其他操作。

7、GoldenDB访问PanWeiDB

GoldenDB分布式（中台中心），访问，PanWeiDB（PanWeiDB版本升级测试库）

8、PanWeiDB访问AntDB

PanWeiDB（PanWeiDB版本升级测试库），访问，AntDB（财务系统）

创建扩展

-- 本地数据库创建表
-- 表1: employees (员工表)
CREATE TABLE employees (
    emp_id int PRIMARY KEY,
    emp_name VARCHAR(50) NOT NULL,
    hire_date DATE,
    salary int,
    department VARCHAR(50),
    resume text,
    photo text
);

-- 表2: departments (部门表)
CREATE TABLE departments (
    dept_id int PRIMARY KEY,
    dept_name VARCHAR(50) UNIQUE,
    manager_id int,
    budget int,
    CONSTRAINT fk_manager FOREIGN KEY (manager_id) REFERENCES employees(emp_id)
);

-- 表3: projects (项目表)
CREATE TABLE projects (
    project_id int PRIMARY KEY,
    project_name VARCHAR(100),
    start_date DATE,
    end_date DATE,
    dept_id int,
    CONSTRAINT fk_dept FOREIGN KEY (dept_id) REFERENCES departments(dept_id)
);

-- 表4: project_assignments (项目分配表)
CREATE TABLE project_assignments (
    assignment_id int PRIMARY KEY,
    emp_id int,
    project_id int,
    role VARCHAR(50),
    CONSTRAINT fk_emp FOREIGN KEY (emp_id) REFERENCES employees(emp_id),
    CONSTRAINT fk_project FOREIGN KEY (project_id) REFERENCES projects(project_id)
);

-- 创建索引
CREATE INDEX idx_emp_dept ON employees(department);
CREATE INDEX idx_proj_dept ON projects(dept_id);
CREATE INDEX idx_assign_role ON project_assignments(role);

插入数据

-- 插入员工数据
DO $$
BEGIN
    FOR i IN 1..1000 LOOP
        INSERT INTO employees VALUES (
            i,
            'Employee' || i,
            CURRENT_DATE - MOD(i, 365),
            5000 + MOD(i, 10000),
            CASE MOD(i, 4)
                WHEN 0 THEN 'IT'
                WHEN 1 THEN 'HR'
                WHEN 2 THEN 'Finance'
                ELSE 'Sales'
            END,
            'aaaaaaaaaa',
            'bbbbbbbbbb'
        );
    END LOOP;
END;
$$;

-- 插入部门数据
INSERT INTO departments VALUES (1, 'IT', 100, 1000000);
INSERT INTO departments VALUES (2, 'HR', 200, 500000);

-- 插入项目数据
INSERT INTO projects VALUES (101, 'ERP System', CURRENT_DATE-100, CURRENT_DATE+200, 1);
INSERT INTO projects VALUES (102, 'HR Portal', CURRENT_DATE-50, CURRENT_DATE+100, 2);

-- 插入项目分配数据
DO $$
BEGIN
    FOR i IN 1..500 LOOP
        INSERT INTO project_assignments VALUES (
        i,
        MOD(i, 1000) + 1,
        CASE MOD(i, 2) WHEN 0 THEN 101 ELSE 102 END,
        CASE MOD(i, 3) WHEN 0 THEN 'Developer' WHEN 1 THEN 'Tester' ELSE 'Manager' END
        );
    END LOOP;
END;
$$;

不支持使用dblink直接访问

testdb1=> SELECT * FROM employees@dblink_antdb WHERE department = 'IT';
ERROR: could not connect to server "dblink_antdb"
DETAIL: expected authentication request from server, but received v
testdb1=>

认证错误

使用外部表的方式访问

create extension postgres_fdw;
grant usage on foreign data wrapper postgres_fdw to testuser1;
create server antdb_fdw_server foreign data wrapper postgres_fdw options(HOST '10.183.103.130',PORT '5432');
create user mapping for testuser1 server antdb_fdw_server options(user 'testuser1',password '3GahEH271!');
create foreign table employees_fdw_antdb (
    emp_id NUMBER(10) ,
    emp_name VARCHAR2(50) NOT NULL,
    hire_date DATE,
    salary NUMBER(10, 2),
    department VARCHAR2(50),
    resume CLOB,
    photo BLOB
) server antdb_fdw_server options(schema_name 'public',table_name 'employees');
create foreign table departments_fdw_antdb (
    dept_id NUMBER(10),
    dept_name VARCHAR2(50) ,
    manager_id NUMBER(10),
    budget NUMBER(15, 2)
) server antdb_fdw_server options(schema_name 'public',table_name 'departments');
create foreign table projects_fdw_antdb (
    project_id NUMBER(10),
    project_name VARCHAR2(100),
    start_date DATE,
    end_date DATE,
    dept_id NUMBER(10)
) server antdb_fdw_server options(schema_name 'public',table_name 'projects');
create foreign table project_assignments_fdw_antdb (
    assignment_id NUMBER(10) ,
    emp_id NUMBER(10),
    project_id NUMBER(10),
    role VARCHAR2(50)
) server antdb_fdw_server options(schema_name 'public',table_name 'project_assignments');

依然无法访问。

9、AntDB访问PanWeiDB

AntDB（财务系统、全面预算系统），访问，PanWeiDB（PanWeiDB版本升级测试库）

create extension postgres_fdw;
grant usage on foreign data wrapper postgres_fdw to testuser1;
create server panweidb_fdw_server foreign data wrapper postgres_fdw options(HOST '10.183.162.150',PORT '17700',dbname 'testdb1');
create user mapping for testuser1 server panweidb_fdw_server options(user 'testuser1',password '3GahEH271');
create foreign table employees_fdw_panweidb (
    emp_id int ,
    emp_name VARCHAR(50) NOT NULL,
    hire_date DATE,
    salary int,
    department VARCHAR(50),
    resume text,
    photo text
) server panweidb_fdw_server options(schema_name 'public',table_name 'employees');

不支持number、varchar2、clob、blob

无法访问外部表

测试用例

以下是一个基于Oracle数据库使用DBLink的详细案例，包含测试用例、完整SQL语句及复杂查询示例：

---

1. 测试用例设计

1.1 创建测试表及约束

-- 本地数据库创建表
-- 表1: employees (员工表)
CREATE TABLE employees (
    emp_id NUMBER(10) PRIMARY KEY,
    emp_name VARCHAR2(50) NOT NULL,
    hire_date DATE,
    salary NUMBER(10, 2),
    department VARCHAR2(50),
    resume CLOB,
    photo BLOB
);

-- 表2: departments (部门表)
CREATE TABLE departments (
    dept_id NUMBER(10) PRIMARY KEY,
    dept_name VARCHAR2(50) UNIQUE,
    manager_id NUMBER(10),
    budget NUMBER(15, 2),
    CONSTRAINT fk_manager FOREIGN KEY (manager_id) REFERENCES employees(emp_id)
);

-- 表3: projects (项目表)
CREATE TABLE projects (
    project_id NUMBER(10) PRIMARY KEY,
    project_name VARCHAR2(100),
    start_date DATE,
    end_date DATE,
    dept_id NUMBER(10),
    CONSTRAINT fk_dept FOREIGN KEY (dept_id) REFERENCES departments(dept_id)
);

-- 表4: project_assignments (项目分配表)
CREATE TABLE project_assignments (
    assignment_id NUMBER(10) PRIMARY KEY,
    emp_id NUMBER(10),
    project_id NUMBER(10),
    role VARCHAR2(50),
    CONSTRAINT fk_emp FOREIGN KEY (emp_id) REFERENCES employees(emp_id),
    CONSTRAINT fk_project FOREIGN KEY (project_id) REFERENCES projects(project_id)
);

-- 创建索引
CREATE INDEX idx_emp_dept ON employees(department);
CREATE INDEX idx_proj_dept ON projects(dept_id);
CREATE INDEX idx_assign_role ON project_assignments(role);

1.2 插入测试数据（通过PL/SQL循环）

-- 插入员工数据
BEGIN
  FOR i IN 1..1000 LOOP
    INSERT INTO employees VALUES (
      i,
      'Employee' || i,
      SYSDATE - MOD(i, 365),
      5000 + MOD(i, 10000),
      CASE MOD(i, 4) WHEN 0 THEN 'IT' WHEN 1 THEN 'HR' WHEN 2 THEN 'Finance' ELSE 'Sales' END,
      EMPTY_CLOB(),
      EMPTY_BLOB()
    );
  END LOOP;
END;
/

-- 插入部门数据
INSERT INTO departments VALUES (1, 'IT', 100, 1000000);
INSERT INTO departments VALUES (2, 'HR', 200, 500000);

-- 插入项目数据
INSERT INTO projects VALUES (101, 'ERP System', SYSDATE-100, SYSDATE+200, 1);
INSERT INTO projects VALUES (102, 'HR Portal', SYSDATE-50, SYSDATE+100, 2);

-- 插入项目分配数据
BEGIN
  FOR i IN 1..500 LOOP
    INSERT INTO project_assignments VALUES (
      i,
      MOD(i, 1000) + 1,
      CASE MOD(i, 2) WHEN 0 THEN 101 ELSE 102 END,
      CASE MOD(i, 3) WHEN 0 THEN 'Developer' WHEN 1 THEN 'Tester' ELSE 'Manager' END
    );
  END LOOP;
END;
/

---

2. DBLink配置与基本SQL

2.1 创建DBLink

-- 创建到远程数据库的DBLink（假设远程数据库名为remote_db）
-- CREATE DATABASE LINK remote_db_link
-- CONNECT TO remote_user IDENTIFIED BY remote_password
-- USING 'remote_db';

grant all on database testdb1 to testuser1;
create extension postgres_fdw;
grant usage on foreign data wrapper postgres_fdw to testuser1;

# 创建到oracle的映射，执行此语句需预先使用 gs_guc generate 命令生成 datasource 文件，此处以如下命令为例：
gs_guc generate -S 'Gs@123456' -D $GAUSSHOME/bin -o usermapping

create database link dblink_pwcrm149 connect to testuser1 identified by '3GahEH271' using postgres_fdw(host '10.183.162.149',port '17700',dbname 'testdb1');

2.2 基本SQL操作

-- 查询远程数据（SELECT）
SELECT * FROM employees@dblink_pwcrm149 WHERE department = 'IT';

-- 插入数据到本地表（INSERT）
INSERT INTO employees
SELECT * FROM employees@dblink_pwcrm149 WHERE emp_id > 500;

-- 更新本地表数据（UPDATE）
UPDATE employees e
SET e.salary = e.salary * 1.1
WHERE EXISTS (
  SELECT dept_id FROM departments@dblink_pwcrm149 d
  WHERE d.dept_name = e.department AND d.dept_name = 'IT'
);

-- 删除数据（DELETE）
DELETE FROM project_assignments
WHERE project_id IN (
  SELECT project_id FROM projects@dblink_pwcrm149 WHERE dept_id = 1
);

---

3. 复杂SQL示例

3.1 多表关联（4表JOIN）

-- 左连接 + 右连接 + 窗口函数
SELECT
    e.emp_name,
    d.dept_name,
    p.project_name,
    pa.role,
    AVG(e.salary) OVER (PARTITION BY d.dept_id) AS avg_dept_salary
FROM employees@dblink_pwcrm149 e
LEFT JOIN departments d ON e.department = d.dept_name
RIGHT JOIN projects@dblink_pwcrm149 p ON d.dept_id = p.dept_id
JOIN project_assignments pa ON pa.project_id = p.project_id
WHERE p.start_date > SYSDATE - 365
AND e.salary > 6000;

3.2 笛卡尔积与过滤条件

-- 笛卡尔积 + CASE判断
SELECT
    e.emp_name,
    p.project_name,
    CASE
        WHEN e.salary > 10000 THEN 'High'
        WHEN e.salary BETWEEN 5000 AND 10000 THEN 'Medium'
        ELSE 'Low'
    END AS salary_level
FROM employees@dblink_pwcrm149 e, projects p
WHERE p.dept_id = 1
AND e.department = 'IT';

3.3 嵌套循环（NESTED LOOP）

-- 强制使用NESTED LOOP（提示方式）
SELECT /*+ USE_NL(e d) */
    e.emp_id,
    d.dept_name
FROM employees e
JOIN departments@dblink_pwcrm149 d ON e.department = d.dept_name
WHERE d.budget > 500000;

3.4 哈希连接（HASH JOIN）

-- 强制使用HASH JOIN
SELECT /*+ USE_HASH(pa p) */
    pa.assignment_id,
    p.project_name
FROM project_assignments pa
JOIN projects@dblink_pwcrm149 p ON pa.project_id = p.project_id
WHERE p.end_date > SYSDATE;

3.5 聚合函数与GROUP BY

-- 聚合函数 + HAVING
SELECT
    d.dept_name,
    COUNT(e.emp_id) AS total_employees,
    SUM(e.salary) AS total_salary
FROM departments@dblink_pwcrm149 d
LEFT JOIN employees e ON d.dept_name = e.department
GROUP BY d.dept_name
HAVING SUM(e.salary) > 100000;

3.6 子查询与EXISTS

-- EXISTS子查询
SELECT
    e.emp_name
FROM employees e
WHERE EXISTS (
    SELECT 1
    FROM project_assignments@dblink_pwcrm149 pa
    WHERE pa.emp_id = e.emp_id
    AND pa.role = 'Manager'
);

3.7 窗口函数与RANK

-- RANK()窗口函数
SELECT
    emp_name,
    salary,
    RANK() OVER (ORDER BY salary DESC) AS salary_rank
FROM employees@dblink_pwcrm149
WHERE department = 'IT';

3.8 复杂条件与函数

-- 字符串函数 + 日期计算
SELECT
    e.emp_name,
    UPPER(d.dept_name) AS dept_upper,
    ADD_MONTHS(e.hire_date, 12) AS hire_anniversary
FROM employees@dblink_pwcrm149 e
JOIN departments d ON e.department = d.dept_name
WHERE TO_CHAR(e.hire_date, 'YYYY') = '2023'
AND LENGTH(e.emp_name) > 5;

总结

• 测试用例：覆盖了多表结构、索引、约束及循环插入数据。
• DBLink操作：实现了跨数据库的增删改查。
• 复杂SQL：包含多表关联、窗口函数、不同连接算法及优化提示。

通过以上案例，可以从多个维度测试不同数据库DBLink的功能性和性能，同时满足复杂业务场景需求。