Question

我有这个查询（Oracle 11g）：

SELECT 
  distinct to_number(val.number_value), val.variant_number
from
  val
left join
  mad on mad.id = val.madid
where 
  val.id = 227753377 and mad.fk_parent =18468
  and (val.language_id = 6003 or val.language_id is null)
  and mad.identifier = 'IDENTIFIER2'
order by val.variant_number

返回给我：

在这种情况下，Oracle不知道该怎么做，因为它引用了所有选定的列，我想。

上面这种情况我需要的是这个结果：

 0 | 1

但在另一个例子中它应该给我：

0 | 1
1 | 2
2 | 3

如果val.number_value是重复的，那么它应该被忽略并为此仅返回1行。怎么做？

Answer 1

如果每个number_value只需要一个variant_number，则可以使用group by和一些汇总功能（例如最小/最大）：

SELECT 
  to_number(val.number_value), min(val.variant_number)
from
  val
left join
  mad on mad.id = val.madid
where 
  val.id = 227753377 and mad.fk_parent =18468
  and (val.language_id = 6003 or val.language_id is null)
  and mad.identifier = 'IDENTIFIER2'
group by to_number(val.number_value)

Answer 2

with numbers as (
  SELECT to_number(val.number_value) as number_value, 
         val.variant_number,
         row_number() over (partition by val.number_value order by val.variant_number) as rn
  from val
    left join mad on mad.id = val.madid
  where val.id = 227753377 and mad.fk_parent =18468
    and (val.language_id = 6003 or val.language_id is null)
    and mad.identifier = 'IDENTIFIER2'
  order by val.variant_number
) 
select number_value, 
       variant_Number
from numbers
where rn = 1;

通过更改窗口功能中的order by，您可以“配置”从重复项中选择哪一行。

Oracle如何在一列上使用distinct

2 个答案: