递归层次汇总查询
在一个数据仓库项目中，人员、组织机构、事实表是这样设计的：
- 组织机构是一个树形结构
- 每一个人员只属于一层组织机构，该层为叶子节点和非叶子节点均可
- 事实表存储每个人员的数据
-- 建立测试表
-- 1. 组织机构
CREATE TABLE org
(
org_id NUMBER,
org_name VARCHAR2 (100),
p_org_id NUMBER
);
-- 2. 职员
CREATE TABLE emp
(
emp_id NUMBER,
emp_name VARCHAR2 (20),
org_id NUMBER
);
-- 3. 事实表
CREATE TABLE fact_data
(
id NUMBER,
data_created DATE,
emp_id NUMBER,
work_hours NUMBER,
task_count NUMBER
);
-- 插入测试数据
-- 组织机构
INSERT INTO ORG (ORG_ID, ORG_NAME, P_ORG_ID)
VALUES (1, '事业部', NULL);
INSERT INTO ORG (ORG_ID, ORG_NAME, P_ORG_ID)
VALUES (2, '技术中心', 1);
INSERT INTO ORG (ORG_ID, ORG_NAME, P_ORG_ID)
VALUES (3, '营销中心', 1);
INSERT INTO ORG (ORG_ID, ORG_NAME, P_ORG_ID)
VALUES (4, '开发组', 2);
INSERT INTO ORG (ORG_ID, ORG_NAME, P_ORG_ID)
VALUES (5, '运维组', 2);
INSERT INTO ORG (ORG_ID, ORG_NAME, P_ORG_ID)
VALUES (6, '行业1', 3);
INSERT INTO ORG (ORG_ID, ORG_NAME, P_ORG_ID)
VALUES (7, '行业2', 3);
INSERT INTO ORG (ORG_ID, ORG_NAME, P_ORG_ID)
VALUES (8, '项目1', 4);
INSERT INTO ORG (ORG_ID, ORG_NAME, P_ORG_ID)
VALUES (9, '项目2', 4);
-- 职员
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (1, '系统管理员', 1);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (2, '技术中心BOSS', 2);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (3, '营销中心BOSS', 3);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (4, '开发组LEADER', 4);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (5, '运维组LEADER', 5);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (6, '开发1', 8);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (7, '开发2', 8);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (8, '开发3', 8);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (9, '开发4', 9);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (10, '开发5', 9);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (11, '开发6', 9);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (12, '销售1', 6);
INSERT INTO EMP (EMP_ID, EMP_NAME, ORG_ID)
VALUES (13, '销售2', 7);
-- 事实表
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (1,
TRUNC (SYSDATE),
1,
10,
3);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (2,
TRUNC (SYSDATE),
2,
6,
15);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (3,
TRUNC (SYSDATE),
3,
4,
7);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (4,
TRUNC (SYSDATE),
4,
12,
5);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (5,
TRUNC (SYSDATE),
5,
30,
13);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (6,
TRUNC (SYSDATE),
6,
11,
3);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (7,
TRUNC (SYSDATE),
7,
5,
13);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (8,
TRUNC (SYSDATE),
8,
6,
6);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (9,
TRUNC (SYSDATE),
9,
1,
2);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (10,
TRUNC (SYSDATE),
10,
14,
4);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (11,
TRUNC (SYSDATE),
11,
11,
11);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (12,
TRUNC (SYSDATE),
12,
8,
6);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (13,
TRUNC (SYSDATE),
13,
9,
5);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (14,
TRUNC (SYSDATE - 1),
1,
10,
3);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (15,
TRUNC (SYSDATE - 1),
2,
6,
15);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (16,
TRUNC (SYSDATE - 1),
3,
4,
7);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (17,
TRUNC (SYSDATE - 1),
4,
12,
5);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (18,
TRUNC (SYSDATE - 1),
5,
30,
13);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (19,
TRUNC (SYSDATE - 1),
6,
11,
3);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (20,
TRUNC (SYSDATE - 1),
7,
5,
13);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (21,
TRUNC (SYSDATE - 1),
8,
6,
6);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (22,
TRUNC (SYSDATE - 1),
9,
1,
2);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (23,
TRUNC (SYSDATE - 1),
10,
14,
4);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (24,
TRUNC (SYSDATE - 1),
11,
11,
11);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (25,
TRUNC (SYSDATE - 1),
12,
8,
6);
INSERT INTO FACT_DATA (ID,
DATA_CREATED,
EMP_ID,
WORK_HOURS,
TASK_COUNT)
VALUES (26,
TRUNC (SYSDATE - 1),
13,
9,
5);
COMMIT;
-- 下钻查询
-- 一级
SELECT f,
root_org_id,
root_org_name,
s_work_hours,
s_task_count
FROM ( SELECT 2 f,
b.root_org_id,
b.root_org_name,
SUM (c.work_hours) s_work_hours,
SUM (c.task_count) s_task_count
FROM emp a,
( SELECT org_id,
CONNECT_BY_ROOT org_id root_org_id,
CONNECT_BY_ROOT org_name root_org_name
FROM org
START WITH p_org_id IS NULL
CONNECT BY PRIOR org_id = p_org_id) b,
fact_data c
WHERE a.org_id = b.org_id AND c.emp_id = a.emp_id
GROUP BY b.root_org_id, b.root_org_name
UNION ALL
SELECT 1,
a.org_id,
a.emp_name,
SUM (c.work_hours),
SUM (c.task_count)
FROM emp a, fact_data c
WHERE org_id IS NULL AND c.emp_id = a.emp_id
GROUP BY a.org_id, a.emp_name)
ORDER BY root_org_id;
-- 二级
SELECT f,
root_org_id,
root_org_name,
s_work_hours,
s_task_count
FROM ( SELECT 2 f,
b.root_org_id,
b.root_org_name,
SUM (c.work_hours) s_work_hours,
SUM (c.task_count) s_task_count
FROM emp a,
( SELECT org_id,
CONNECT_BY_ROOT org_id root_org_id,
CONNECT_BY_ROOT org_name root_org_name
FROM org
START WITH p_org_id = 1
CONNECT BY PRIOR org_id = p_org_id) b,
fact_data c
WHERE a.org_id = b.org_id AND c.emp_id = a.emp_id
GROUP BY b.root_org_id, b.root_org_name
UNION ALL
SELECT 1,
a.org_id,
a.emp_name,
SUM (c.work_hours),
SUM (c.task_count)
FROM emp a, fact_data c
WHERE org_id = 1 AND c.emp_id = a.emp_id
GROUP BY a.org_id, a.emp_name)
ORDER BY root_org_id;
-- 三级
SELECT f,
root_org_id,
root_org_name,
s_work_hours,
s_task_count
FROM ( SELECT 2 f,
b.root_org_id,
b.root_org_name,
SUM (c.work_hours) s_work_hours,
SUM (c.task_count) s_task_count
FROM emp a,
( SELECT org_id,
CONNECT_BY_ROOT org_id root_org_id,
CONNECT_BY_ROOT org_name root_org_name
FROM org
START WITH p_org_id = 2
CONNECT BY PRIOR org_id = p_org_id) b,
fact_data c
WHERE a.org_id = b.org_id AND c.emp_id = a.emp_id
GROUP BY b.root_org_id, b.root_org_name
UNION ALL
SELECT 1,
a.org_id,
a.emp_name,
SUM (c.work_hours),
SUM (c.task_count)
FROM emp a, fact_data c
WHERE org_id = 2 AND c.emp_id = a.emp_id
GROUP BY a.org_id, a.emp_name)
ORDER BY root_org_id;
-- 四级
SELECT f,
root_org_id,
root_org_name,
s_work_hours,
s_task_count
FROM ( SELECT 2 f,
b.root_org_id,
b.root_org_name,
SUM (c.work_hours) s_work_hours,
SUM (c.task_count) s_task_count
FROM emp a,
( SELECT org_id,
CONNECT_BY_ROOT org_id root_org_id,
CONNECT_BY_ROOT org_name root_org_name
FROM org
START WITH p_org_id = 4
CONNECT BY PRIOR org_id = p_org_id) b,
fact_data c
WHERE a.org_id = b.org_id AND c.emp_id = a.emp_id
GROUP BY b.root_org_id, b.root_org_name
UNION ALL
SELECT 1,
a.org_id,
a.emp_name,
SUM (c.work_hours),
SUM (c.task_count)
FROM emp a, fact_data c
WHERE org_id = 4 AND c.emp_id = a.emp_id
GROUP BY a.org_id, a.emp_name)
ORDER BY root_org_id; 在这个实现中通过传入上级组织机构ID实现下钻汇总查询。