计算列上的T-SQL列别名 - 无效的列名称

Question

我正在使用别名来引用计算列.这是我正在努力工作的实际代码的片段,用于计算相似性并返回相似性得分为3或更高的匹配.

select [FirstName], difference([FirstName], 'mitch') as similarity
from [Dev].[dbo].[Name]
where similarity > 2
order by similarity desc

错误:

select [FirstName], difference([FirstName], 'mitch') as similarity
from [Dev].[dbo].[Name]
where similarity > 2
order by similarity desc

由于相似性不是一个真正的专栏,我将如何使这项工作？

Answer 1

列别名和计算SELECT在查询的projection()阶段执行,该阶段发生在选择(WHERE和JOIN)阶段之后.因此,它们不能在WHERE条款或JOIN条件中引用,因为它们尚不存在.您可以将查询与SELECT子句一起用作子查询,也可以在WHERE子句中复制计算:

select * 

from
(select [FirstName], difference([FirstName], 'mitch') as similarity
from [Dev].[dbo].[Name]) src

where similarity > 2
order by similarity desc

要么

select [FirstName], difference([FirstName], 'mitch') as similarity
from [Dev].[dbo].[Name]
where difference([FirstName], 'mitch') > 2
order by similarity desc

Answer 2

所有答案都可以解决您的问题,但对于复杂的情况,您无法复制查询.

正确的方法是使用CROSS和APPLY

select [FirstName], similarity
from [Dev].[dbo].[Name]
cross apply
  (
     select similarity =
     difference([FirstName], 'mitch')  
  )computed_column
where similarity > 2
order by similarity desc

whit CROSS和APPLY你可以在查询的任何地方使用你的计算列

Answer 3

尝试:

SELECT * 
  FROM (
        SELECT [FirstName], difference([FirstName], 'mitch') as similarity
            FROM [Dev].[dbo].[Name]
        ) a
WHERE similarity > 2
ORDER BY similarity desc