我尝试了几种替代方案，但由于我的数据记录有重复的值，ROW_NUMBER版本似乎不是我的选择。这里是我使用的查询(NTILE版本):

SELECT distinct
   CustomerId,
   (
       MAX(CASE WHEN Percent50_Asc=1 THEN TotalDue END) OVER (PARTITION BY CustomerId)  +
       MIN(CASE WHEN Percent50_desc=1 THEN TotalDue END) OVER (PARTITION BY CustomerId) 
   )/2 MEDIAN
FROM
(
   SELECT
      CustomerId,
      TotalDue,
     NTILE(2) OVER (
         PARTITION BY CustomerId
         ORDER BY TotalDue ASC) AS Percent50_Asc,
     NTILE(2) OVER (
         PARTITION BY CustomerId
         ORDER BY TotalDue DESC) AS Percent50_desc
   FROM Sales.SalesOrderHeader SOH
) x
ORDER BY CustomerId;

从员工表中得到工资的中位数

with cte as (select salary, ROW_NUMBER() over (order by salary asc) as num from employees)

select avg(salary) from cte where num in ((select (count(*)+1)/2 from employees), (select (count(*)+2)/2 from employees));

这是我能想到的求中位数的最优解。示例中的名称基于Justin示例。确保表有索引 销售。SalesOrderHeader以索引列CustomerId和TotalDue的顺序存在。

SELECT
 sohCount.CustomerId,
 AVG(sohMid.TotalDue) as TotalDueMedian
FROM 
(SELECT 
  soh.CustomerId,
  COUNT(*) as NumberOfRows
FROM 
  Sales.SalesOrderHeader soh 
GROUP BY soh.CustomerId) As sohCount
CROSS APPLY 
    (Select 
       soh.TotalDue
    FROM 
    Sales.SalesOrderHeader soh 
    WHERE soh.CustomerId = sohCount.CustomerId 
    ORDER BY soh.TotalDue
    OFFSET sohCount.NumberOfRows / 2 - ((sohCount.NumberOfRows + 1) % 2) ROWS 
    FETCH NEXT 1 + ((sohCount.NumberOfRows + 1) % 2) ROWS ONLY
    ) As sohMid
GROUP BY sohCount.CustomerId

更新

我有点不确定哪种方法性能最好，所以我比较了我的方法Justin Grants和Jeff Atwoods，在一个批量中运行基于这三种方法的查询，每个查询的批量成本为:

没有指数:

我的30% Justin Grants 13% Jeff Atwoods 58%

还有index

我的3%。 Justin Grants 10% Jeff Atwoods 87%

I tried to see how well the queries scale if you have index by creating more data from around 14 000 rows by a factor of 2 up to 512 which means in the end around 7,2 millions rows. Note I made sure CustomeId field where unique for each time I did a single copy, so the proportion of rows compared to unique instance of CustomerId was kept constant. While I was doing this I ran executions where I rebuilt index afterwards, and I noticed the results stabilized at around a factor of 128 with the data I had to these values:

我的3%。 贾斯汀·格兰特5% Jeff Atwoods 92%

我想知道，在保持惟一CustomerId不变的情况下，扩展行数会如何影响性能，因此我设置了一个新的测试，在其中执行了上述操作。现在，批成本比率并没有稳定下来，而是不断分化，每个CustomerId平均大约有20行，最后每个这样唯一的Id大约有10000行。数字如下:

我的4% 贾斯汀60% 杰夫斯35%

通过比较结果，我确保我正确地实现了每个方法。 我的结论是，只要索引存在，我使用的方法通常更快。还要注意，本文针对这个特定问题推荐使用这种方法https://www.microsoftpressstore.com/articles/article.aspx?p=2314819&seqNum=5

进一步提高对该查询的后续调用的性能的一种方法是在辅助表中持久化计数信息。您甚至可以通过一个触发器来维护它，该触发器更新并保存有关依赖于CustomerId的SalesOrderHeader行计数的信息，当然您也可以简单地存储中值。

以下解决方案在这些假设下有效:

无重复值 没有取消

代码:

IF OBJECT_ID('dbo.R', 'U') IS NOT NULL
  DROP TABLE dbo.R

CREATE TABLE R (
    A FLOAT NOT NULL);

INSERT INTO R VALUES (1);
INSERT INTO R VALUES (2);
INSERT INTO R VALUES (3);
INSERT INTO R VALUES (4);
INSERT INTO R VALUES (5);
INSERT INTO R VALUES (6);

-- Returns Median(R)
select SUM(A) / CAST(COUNT(A) AS FLOAT)
from R R1 
where ((select count(A) from R R2 where R1.A > R2.A) = 
      (select count(A) from R R2 where R1.A < R2.A)) OR
      ((select count(A) from R R2 where R1.A > R2.A) + 1 = 
      (select count(A) from R R2 where R1.A < R2.A)) OR
      ((select count(A) from R R2 where R1.A > R2.A) = 
      (select count(A) from R R2 where R1.A < R2.A) + 1) ; 

以下是我的解决方案:

with tempa as

 (

    select value,row_number() over (order by value) as Rn,/* Assigning a 
                                                           row_number */
           count(value) over () as Cnt /*Taking total count of the values */
    from numbers
    where value is not null /* Excluding the null values */
 ),

tempb as

  (

    /* Since we don't know whether the number of rows is odd or even, we shall 
     consider both the scenarios */

    select round(cnt/2) as Ref from tempa where mod(cnt,2)=1
    union all
    select round(cnt/2) a Ref from tempa where mod(cnt,2)=0
     union all
    select round(cnt/2) + 1 as Ref from tempa where mod(cnt,2)=0
   )
  select avg(value) as Median_Value

  from tempa where rn in

    ( select Ref from tempb);

下面的查询从一列中的值列表返回中位数。它不能作为聚合函数使用，也不能与聚合函数一起使用，但仍然可以在内部select中使用带有WHERE子句的子查询。

SQL Server 2005+:

SELECT TOP 1 value from
(
    SELECT TOP 50 PERCENT value 
    FROM table_name 
    ORDER BY  value
)for_median
ORDER BY value DESC