用CSV计数使用SQL Server 2008 R2

Question

我想统计在csv列下的记录数量，并且也希望显示它们不同。

例：

表：

create table test1 
(
cola int, 
colb varchar(10) 
) 

一些记录：

insert into test1 values(1,'SAM'); 
insert into test1 values(2,'SAM'); 
insert into test1 values(3,'JACK'); 
insert into test1 values(4,'MAK'); 

脚本：

with cte (colb) 
AS 
(
    SELECT SUBSTRING(
    (SELECT ',' + s.colb 
    FROM test1 as s 
    ORDER BY s.colb 
    FOR XML PATH('')),2,200) AS CSV 
) 
select count(*) as NumberOfNames,colb as Names 
from cte 
group by colb 

结果：

NumberOfNames   Names 
----------------------------------- 
    1   JACK,MAK,SAM,SAM 

预期结果：

NumberOfNames   Names 
----------------------------------- 
    3    JACK,MAK,SAM 

Answer 1

你的CTE只有串接名称，也不会指望他们。 根据您的例子中，查询可能是这样的：

SELECT COUNT(DISTINCT colb) AS NumberOfNames, 
(SELECT STUFF((SELECT ',' + s.colb FROM test1 as s ORDER BY s.colb FOR XML PATH('')), 1, 1, '')) AS CSV 
FROM test1 

结果：

NumberOfNames CSV 
------------- ---------------- 
3    JACK,MAK,SAM,SAM 

在使用STUFF函数删除第一个逗号。

Answer 2

这个怎么样....

Declare @tempVariable as Varchar(100) 
Select @tempVariable= coalesce(@tempVariable+',','')+ colb 
from 
(
    Select distinct colb from test1 
) as te 
Select len(@tempVariable) - len(replace(@tempVariable, ',', ''))+1, @tempVariable 

结果：

​​