Question

我有下面的SQL查询。我正在使用oracle 10g。在这里，我结合了两个查询的结果。

select distinct combined.some_id from (
  SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H')
        union all  
        SELECT DISTINCT e.some_id FROM tableA_Replica e where
        e.some_code  ='ABC' and e.code_two IN('S','H') and e.some_id NOT IN ( SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H'))
    ) combined,tableA_Replica_join_table_ONE x where 
        combined.some_id = x.some_id(+)
        AND (x.status   IN('ACTIVE','INACTIVE'))


        UNION

select distinct combined.some_id from (
  SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H')
        union all  
        SELECT DISTINCT e.some_id FROM tableA_Replica e where
        e.some_code  ='ABC' and e.code_two IN('S','H') and e.some_id NOT IN ( SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H'))
    ) combined,tableA_Replica_join_table_TWO x where 
        combined.some_id = x.some_id(+)
        AND (x.status   IN('ACTIVE','INACTIVE'))

下面两个查询中的

是常见的。

SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H')
        union all  
        SELECT DISTINCT e.some_id FROM tableA_Replica e where
        e.some_code  ='ABC' and e.code_two IN('S','H') and e.some_id NOT IN ( SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H'))

如何在两个查询中避免代码重复？

Answer 1

此查询删除了大量重复内容：

select distinct combined.some_id from (
        SELECT distinct l.some_id FROM tableA
           union
        SELECT distinct e.some_id FROM tableA_Replica
    ) combined
    inner join (
        select x.status, x.some_id from tableA_Replica_join_table_ONE x 
             union
        select y.status, y.some_id from tableA_Replica_join_table_TWO y 
    ) join_table
    on combined.some_id = join_table.some_id(+) and
       join_table.status IN('ACTIVE','INACTIVE') and
       combined.some_code  ='ABC' and l.code_two IN('S','H')

最好的查询总是取决于数据的结构。

你绝对不需要做的一件事：

SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H')
        union all  
        SELECT DISTINCT e.some_id FROM tableA_Replica e where
        e.some_code  ='ABC' and e.code_two IN('S','H') and e.some_id NOT IN ( SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H'))

第二个查询中不需要NOT IN子句。只需使用union代替union all - 它会自动删除重复项。

需要考虑的另一件事是：您在很多地方使用distinct。你真的需要所有这些吗？有些人似乎到处都使用distinct来防止重复。但是，这是低效的，如果您不确切知道它在做什么，可能会导致细微的错误。

通常，只有在您明确决定需要从特定查询中删除重复项时才使用distinct。（如果没有看到你的数据，就不可能知道你是否需要所有这些数据，但是大量数据让我产生怀疑）。

Answer 2

您可以使用公用表表达式（CTE，有时也称为“WITH语句”）：

WITH combined AS (
    SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H')
        union all  
        SELECT DISTINCT e.some_id FROM tableA_Replica e where
        e.some_code  ='ABC' and e.code_two IN('S','H') and e.some_id NOT IN ( SELECT DISTINCT l.some_id FROM tableA l where
        l.some_code  ='ABC' and l.code_two IN('S','H'))
)
SELECT distinct combined.some_id
FROM combined, tableA_Replica_join_table_ONE x
WHERE combined.some_id = x.some_id(+)
      AND (x.status   IN('ACTIVE','INACTIVE'))

UNION

SELECT distinct combined.some_id
FROM combined, tableA_Replica_join_table_TWO x
WHERE combined.some_id = x.some_id(+)
      AND (x.status   IN('ACTIVE','INACTIVE'))

删除sql查询中的重复？

2 个答案: