XML 上的 DISTINCT 与 CAST

Question

我有 4 个 table。 table 不包含任何 foreign key 引用，其中名称的类型为 xml

t1:

| id  | name           |
+-----+----------------+
| a1  | <ID>a1<ID>     |
| a2  | <ID>a2<ID>     |
| a3  | <ID>a3<ID>     |

t2:

| id  | name        |   t1_id      |
+-----+-------------+--------------+
| b1  | <ID>b1<ID>  | a1           |
| b2  | <ID>b2<ID>  | a2           |
| b3  | <ID>b3<ID>  | a3           |
| b4  | <ID>b4<ID>  | a1           |

t3:

| id  | name        |   t2_id      |
+-----+-------------+--------------+
| c1  | <ID>c1<ID>  | b1           |
| c2  | <ID>c2<ID>  | b2           |
| c3  | <ID>c3<ID>  | b3           |

t4:

| id  | name        |   t3_id      |
+-----+-------------+--------------+
| d1  | <ID>d1<ID>  | c2           |
| d2  | <ID>d2<ID>  | c3           |

我需要获取所有相关数据。当我指定 'a1' 时，我期望的是

| name         | name        |   name       |   name       |
+--------------+-------------+--------------+--------------+
| <ID>a1<ID>   | <ID>b1<ID>  | <ID>c1<ID>   |              |
|              | <ID>b4<ID>  |              |              |

得到的结果：

| name         | name        |   name       |   name       |
+--------------+-------------+--------------+--------------+
| <ID>a1<ID>   | <ID>b1<ID>  | <ID>c1<ID>   |              |
| <ID>a1<ID>   | <ID>b4<ID>  |              |              |

为了解决这个问题，我使用了

SELECT DISTINCT CAST (a.name AS TEXT), CAST (b.name AS TEXT), CAST (c.name AS TEXT), CAST (d.name AS TEXT)
FROM t1 AS a

LEFT JOIN t2 AS b
ON b.id = 
CAST((xpath('/ID/text()', b.name))[1] AS TEXT) 

LEFT JOIN t3 AS c
ON c.id = 
CAST((xpath('/ID/text()', c.name))[1] AS TEXT) 

LEFT JOIN t4 AS d
ON d.id = 
CAST((xpath('/ID/text()', d.name))[1] AS TEXT)

以上方法无效。如何解决这个问题。这是优化查询吗？我试过嵌套查询。 Distinct 问题也出现在那里

实现此目标的最佳优化方法是什么？使用 JOIN 或 Nested query?

Answer 1

您显示的查询与 table 示例中的逻辑不匹配。这将：

SELECT t1.name::text, t2.name::text, t3.name::text, t4.name::text
FROM   t1
LEFT   JOIN t2 b ON t2.t1_id = t1.id
LEFT   JOIN t3 c ON t3.t2_id = t2.id
LEFT   JOIN t4 d ON t4.t3_id = t3.id
WHERE  t1.id = 'a1';

而且您的问题中没有任何内容可以保证使用 DISTINCT。

如果您真的想在显示的有序列表的同一列中隐藏重复的名称：

SELECT CASE WHEN rn1 = 1 THEN name1 END AS name1
     , CASE WHEN rn2 = 1 THEN name2 END AS name2
     , CASE WHEN rn3 = 1 THEN name3 END AS name3
     , CASE WHEN rn4 = 1 THEN name4 END AS name4
FROM  (
   SELECT t1.name::text AS name1, t2.name::text AS name2, t3.name::text AS name3, t4.name::text AS name4
        , row_number() OVER (PARTITION BY t1.id ORDER ORDER BY t2.id, t3.id, t4.id) AS rn1
        , row_number() OVER (PARTITION BY t1.id, t2.id ORDER ORDER BY t3.id, t4.id) AS rn2
        , row_number() OVER (PARTITION BY t1.id, t2.id, t3.id ORDER ORDER BY t4.id) AS rn3
   FROM   t1
   LEFT   JOIN t2 b ON t2.t1_id = t1.id
   LEFT   JOIN t3 c ON t3.t2_id = t2.id
   LEFT   JOIN t4 d ON t4.t3_id = t3.id
   WHERE  t1.id = 'a1'
   ) sub
ORDER BY t1.id, t2.id, t3.id, t4.id;

XML 上的 DISTINCT 与 CAST

DISTINCT on XML with CAST

sql

postgresql

join

distinct