Oracle - 在分层查询中使用先连接并将它们放入行中答案

【问题标题】：Oracle - using connect by prior in Hierarchical Queries and put them into rowsOracle - 在分层查询中使用先连接并将它们放入行中
【发布时间】：2021-11-05 01:03:49
【问题描述】：

我是 Oracle 的新手，我将在实现平面表而不是分层表之前使用 connect by。但我有点困惑。我的桌子是这样的：员工表：

empID	empName	managerID
100	Sara	110
101	Ben	111
102	Alex	110
110	Ross	111
111	Mon	NULL

我要这样改变表格（输出）：

emp	empName	subBoss	subBossName	Boss	BossName
100	Sara	110	Ross	111	Mon
101	Ben	111	Mon	NULL	NULL
102	Alex	110	Ross	111	Mon
110	Ross	111	Mon	NULL	NULL
111	Mon	NULL	NULL	NULL	NULL

【问题讨论】：

标签： sql oracle connect-by

【解决方案1】：

MDO 答案的简短版本是 -

WITH DATA AS (SELECT 100 AS empID, 'Sara' AS empName, 110 AS managerID FROM DUAL UNION ALL
              SELECT 101, 'Ben', 111 FROM DUAL UNION ALL
              SELECT 102, 'Alex', 110 FROM DUAL UNION ALL
              SELECT 110, 'Ross', 111 FROM DUAL UNION ALL
              SELECT 111, 'Mon', NULL FROM DUAL)
 SELECT empID emp, empName, prior empID subBoss, prior empName subBossName
    , prior t.managerID Boss
    ,CASE WHEN PRIOR t.managerID IS NOT NULL THEN CONNECT_BY_ROOT(t.empName) END AS BossName
  FROM DATA T
 START WITH managerID IS NULL
CONNECT BY PRIOR empID = managerID
  ORDER BY 1;

Demo.

【讨论】：

如果层次结构有 4 个（或更多）级别，那么这将从层次结构的不同级别 db<>fiddle 获取 boss 的值 id 和 name。

【解决方案2】：

您也可以为此使用 connect by 子句。

select empID, empName
  , subBoss, subBossName
  , (select boss.empID from YourTable boss where boss.empID = subBossManagerID) Boss
  , (select boss.empName from YourTable boss where boss.empID = subBossManagerID) BossName
from (
  select empID, empName
    , prior empID subBoss, prior empName subBossName
    , prior t.managerID subBossManagerID
  from YourTable t
  start with managerID is null
  connect by prior empID = managerID
)
order by 1
;

test here

【讨论】：

很高兴看到我能提供帮助。

【解决方案3】：

您可以为此使用递归子查询分解子句：

WITH hierarchy (empID, empName, subBoss, subBossName, boss, bossName, depth, managerId) AS (
  SELECT empID,
         empName,
         CAST(NULL AS NUMBER),
         CAST(NULL AS VARCHAR2(20)),
         CAST(NULL AS NUMBER),
         CAST(NULL AS VARCHAR2(20)),
         1,
         managerID
  FROM   empTbl
UNION ALL
  SELECT h.empID,
         h.empName,
         CASE h.depth
         WHEN 1 THEN e.empID
         ELSE h.subBoss
         END,
         CASE h.depth
         WHEN 1 THEN e.empName
         ELSE h.subBossName
         END,
         CASE h.depth
         WHEN 2 THEN e.empID
         ELSE h.boss
         END,
         CASE h.depth
         WHEN 2 THEN e.empName
         ELSE h.bossName
         END,
         h.depth + 1,
         e.managerID
  FROM   hierarchy h
         LEFT OUTER JOIN empTbl e
         ON (h.managerID = e.empID)
  WHERE  depth < 3
)
CYCLE empID, depth SET is_cycle TO 1 DEFAULT 0
SELECT empID, empName, subBoss, subBossName, boss, bossName
FROM   hierarchy
WHERE  depth = 3;

或者，您可以使用分层查询和数据透视：

SELECT emp_id AS empID,
       emp_name AS empName,
       subboss_id AS subbossid,
       subboss_name AS subbossname,
       boss_id AS bossid,
       boss_name AS bossname
FROM   (
  SELECT CONNECT_BY_ROOT(empID) AS root_empid,
         empID,
         empName,
         LEVEL AS depth
  FROM   empTbl
  WHERE  LEVEL <= 3
  CONNECT BY PRIOR managerID = empID
)
PIVOT (
  MAX(empID) AS id,
  MAX(empName) AS name
  FOR depth IN (
    1 AS emp,
    2 AS subBoss,
    3 AS boss
  )
)
ORDER BY empid;

其中，对于样本数据：

CREATE TABLE empTbl (empID, empName, managerID) AS
SELECT 100, 'Sara', 110 FROM DUAL UNION ALL
SELECT 101, 'Ben',  111 FROM DUAL UNION ALL
SELECT 102, 'Alex', 110 FROM DUAL UNION ALL
SELECT 110, 'Ross', 111 FROM DUAL UNION ALL
SELECT 111, 'Mon',  NULL FROM DUAL;

两个输出：

EMPID EMPNAME SUBBOSS SUBBOSSNAME BOSS BOSSNAME

100 Sara 110 Ross 111 Mon

102 Alex 110 Ross 111 Mon

101 Ben 111 Mon

110 Ross 111 Mon

111 Mon

db小提琴here

【讨论】：