Question

我对SQL没有最好的理解，尽力学习。我已经构建了两个单独的查询，我希望从查询1中获取结果并将其除以查询2.

查询1（结果为114）

select count(*) as Numerator

from (

SELECT   *,  
           DATEDIFF(SECOND, xx, yy) AS SecondDiff, 
           DATEDIFF(DAY, xx, yy) AS DayDiff

FROM database1.dbo.tbl1

WHERE  month(datecompleted) = month(dateadd(month,-1,current_timestamp))
       and year(datecompleted) = year(dateadd(month,-1,current_timestamp))
       and DATEDIFF(DAY, xx,yy) <= 15

)t1

查询2（结果为131）

select count(*) as Denominator

from(

SELECT   *,  
           DATEDIFF(SECOND, xx, yy) AS SecondDiff, 
           DATEDIFF(DAY, xx, yy) AS DayDiff

FROM database1.dbo.tbl1

WHERE  month(datecompleted) = month(dateadd(month,-1,current_timestamp))
       and year(datecompleted) = year(dateadd(month,-1,current_timestamp))


)t2

现在我如何从查询1中获取结果并将其除以查询2并将最终结果设为（列名：Total_X）？我曾尝试使用union，cross join，cross apply但我似乎无法获得有效的查询！

非常感谢每一条提示和建议

由于

Answer 1

我认为你正在寻找这个：

SELECT AVG(CASE WHEN DATEDIFF(DAY, xx,yy) <= 15 THEN 1.0 ELSE 0 END)
FROM database1.dbo.tbl1
WHERE month(datecompleted) = month(dateadd(month,-1,current_timestamp)) AND
      year(datecompleted) = year(dateadd(month,-1,current_timestamp));

您还可以将WHERE子句简化为：

SELECT AVG(CASE WHEN DATEDIFF(DAY, xx,yy) <= 15 THEN 1.0 ELSE 0 END)
FROM database1.dbo.tbl1
WHERE datecompleted < DATEADD(day, 1 - DAY(current_timestamp), CONVERT(DATE, datecompleted)) AND
      datecompleted >= DATEADD(month, 1, DATEADD(day, 1 - DAY(current_timestamp), CONVERT(DATE, datecompleted))) AND

此版本可以使用datecompleted上的索引。

Answer 2

您可以在两个查询之间使用CROSS JOIN，例如：

  select TT1.Numerator/ TT2.Denominator
  from (
    select count(*) as Numerator

  from (

  SELECT   *,  
             DATEDIFF(SECOND, xx, yy) AS SecondDiff, 
             DATEDIFF(DAY, xx, yy) AS DayDiff

  FROM database1.dbo.tbl1

  WHERE  month(datecompleted) = month(dateadd(month,-1,current_timestamp))
         and year(datecompleted) = year(dateadd(month,-1,current_timestamp))
         and DATEDIFF(DAY, xx,yy) <= 15

  )t1) TT1 
  CROSS JOIN ( 

  select count(*) as Denominator

  from(

  SELECT   *,  
             DATEDIFF(SECOND, xx, yy) AS SecondDiff, 
             DATEDIFF(DAY, xx, yy) AS DayDiff

  FROM database1.dbo.tbl1

  WHERE  month(datecompleted) = month(dateadd(month,-1,current_timestamp))
         and year(datecompleted) = year(dateadd(month,-1,current_timestamp))


  )t2 ) TT2

可能与整数除法有关，尝试简单的转换

select cast(TT1.Numerator AS DECIMAL(9,2))/CAST( TT2.Denominator AS DECIMAL(9.2))
  from (
    select count(*) as Numerator

  from (

  SELECT   *,  
             DATEDIFF(SECOND, xx, yy) AS SecondDiff, 
             DATEDIFF(DAY, xx, yy) AS DayDiff

  FROM database1.dbo.tbl1

  WHERE  month(datecompleted) = month(dateadd(month,-1,current_timestamp))
         and year(datecompleted) = year(dateadd(month,-1,current_timestamp))
         and DATEDIFF(DAY, xx,yy) <= 15

  )t1) TT1 
  CROSS JOIN ( 

  select count(*) as Denominator

  from(

  SELECT   *,  
             DATEDIFF(SECOND, xx, yy) AS SecondDiff, 
             DATEDIFF(DAY, xx, yy) AS DayDiff

  FROM database1.dbo.tbl1

  WHERE  month(datecompleted) = month(dateadd(month,-1,current_timestamp))
         and year(datecompleted) = year(dateadd(month,-1,current_timestamp))


  )t2 ) TT2

Answer 3

您可以将行转换为列并进行分割。假设你的第一个查询是从双重中选择1个值，'nm'类型; 第二个查询是从双

将rwos转换为列并除以

选择sum（decode（type，'nm'，value））分子，
  sum（decode（type，'dm'，value））分母，   总和（解码（类型， '纳米'，值））/ SUM（解码（类型， 'DM'，值））   divide_value from（选择1值，'nm'类型来自双联盟
  从1,2
中选择2个值，'dm'类型来自双重）

或直接

  select 1/(select 2 from dual) from dual;

Answer 4

另一种方法是使用计算表表达式。

with numerator
as (
select count(*) n1
from someTable
) 
, denominator as
(select count(*) n2
from anotherTable)

select n1 / n2
from numerator, denominator
where n2 > 0

请注意

from numerator, denominator

实际上是一个交叉连接，正如@scaisEdge

所建议的那样

另请注意，您的实际查询看起来比必要的更复杂。

Sql - 如何计算2个单独的子查询？

4 个答案: