我有一个查询,从下面的表
创建一个H_TREE列+------+--------+
| UNIT | M_UNIT |
+------+--------+
| 10 | 12 |
| 15 | 19 |
| 12 | 16 |
| 13 | 15 |
| 19 | 14 |
| 14 | 11 |
+------+--------+
以下是我正在使用的查询,
WITH data (unit, m_unit) AS (
SELECT 10, 12 FROM dual UNION ALL
SELECT 15, 19 FROM dual UNION ALL
SELECT 12, 16 FROM dual UNION ALL
SELECT 13, 15 FROM dual UNION ALL
SELECT 19, 14 FROM dual UNION ALL
SELECT 14, 11 FROM dual)
SELECT
unit,
m_unit,
unit || ',' || listagg(root_unit, ',') WITHIN GROUP (ORDER BY depth) h_tree
FROM (
SELECT
id, unit, m_unit,
LEVEL depth, CONNECT_BY_ROOT m_unit root_unit
FROM
(SELECT ROWNUM id, unit, m_unit FROM data) data
CONNECT BY
PRIOR unit = m_unit)
GROUP BY
id,
unit,
m_unit
上面的查询返回
+------+--------+----------------+
| UNIT | M_UNIT | H_TREE |
+------+--------+----------------+
| 10 | 12 | 10,12,16 |
| 15 | 19 | 15,19,14,11 |
| 12 | 16 | 12,16 |
| 13 | 15 | 13,15,19,14,11 |
| 19 | 14 | 19,14,11 |
| 14 | 11 | 14,11 |
+------+--------+----------------+
我想颠倒H_TREE列的顺序,以便它如下所示,
+------+--------+----------------+
| UNIT | M_UNIT | H_TREE |
+------+--------+----------------+
| 10 | 12 | 16,12,10 |
| 15 | 19 | 11,14,19,15 |
| 12 | 16 | 16,12 |
| 13 | 15 | 11,14,19,15,13 |
| 19 | 14 | 11,14,19 |
| 14 | 11 | 11,14 |
+------+--------+----------------+
根据评论:WITHIN GROUP(ORDER BY depth DESC)给出了以下结果,
+------+--------+----------------+
| UNIT | M_UNIT | H_TREE |
+------+--------+----------------+
| 10 | 12 | 10,16,12 |
| 15 | 19 | 15,11,14,19 |
| 12 | 16 | 12,16 |
| 13 | 15 | 13,11,14,19,15|
| 19 | 14 | 19,11,14 |
| 14 | 11 | 14,11 |
+------+--------+----------------+
答案 0 :(得分:1)
尝试:
WITH data (unit, m_unit) AS (
SELECT 10, 12 FROM dual UNION ALL
SELECT 15, 19 FROM dual UNION ALL
SELECT 12, 16 FROM dual UNION ALL
SELECT 13, 15 FROM dual UNION ALL
SELECT 19, 14 FROM dual UNION ALL
SELECT 14, 11 FROM dual)
SELECT
unit,
m_unit,
listagg(root_unit, ',') WITHIN GROUP (ORDER BY depth desc) || ',' || unit as h_tree
FROM (
SELECT
id, unit, m_unit,
LEVEL depth, CONNECT_BY_ROOT m_unit root_unit
FROM
(SELECT ROWNUM id, unit, m_unit FROM data) data
CONNECT BY
PRIOR unit = m_unit)
GROUP BY
id,
unit,
m_unit