将站点列表和出现次数连接到一列中答案

【问题标题】：Concatenating a List of of sites and count of occurrences into a column将站点列表和出现次数连接到一列中
【发布时间】：2015-01-07 06:55:29
【问题描述】：

使用此页面：Can I Comma Delimit Multiple Rows Into One Column?

...我已经能够得出这样的结果：

使用这个查询：

SELECT
    [Client ID],
    STUFF((SELECT ', ' + [Location (counts)]
           FROM (
                SELECT TOP 100 PERCENT [Client ID]
                    , ltrim(str([Store Num])) + ' (' + ltrim(str(count([Store Num]))) + ') ' [Location (counts)]
                FROM @tbl_coreData
                GROUP BY [Store Num], [Client ID]
                ORDER BY COUNT([Store Num]) DESC) tblThis 
           WHERE ([Store Num] = Result.[Store Num] 
             AND [Client ID] = Result.[Client ID]) 
           FOR XML PATH ('')),1,1,'') AS BATCHNOLIST
FROM @tbl_coreData AS Result
GROUP BY [Client ID], [Store Num]

我希望列表按出现次数（括号中的数字）排序。我的尝试是

SELECT TOP 100 PERCENT 
    [Client ID]

和

ORDER BY COUNT([Store Num]) DESC

但这不起作用。

如何使连接的值按出现的降序显示？

一如既往，欢迎并感谢所有帮助。

【问题讨论】：

标签： sql sql-server tsql subquery concatenation

【解决方案1】：

经过一番烦躁和使用霍根的建议，我想出了这个：

DECLARE @Heroes TABLE (
    [HeroName]      VARCHAR(20),
    [HeroFriend]    VARCHAR(20)
)

INSERT INTO @Heroes ( [HeroName], [HeroFriend] )
VALUES  ( 'Superman', 'Aquaman' ), 
        ( 'Superman', 'Batman' ), 
        ( 'Superman', 'Wolverine' ), 
        ( 'Superman', 'Batman' ), 
        ( 'Superman', 'Wolverine' ), 
        ( 'Superman', 'Batman' ), 
        ( 'Superman', 'Ironman' ), 
        ( 'Superman', 'Wolverine' ), 
        ( 'Superman', 'Batman' ), 
        ( 'Superman', 'Ironman' ), 
        ( 'Superman', 'Wolverine' ), 
        ( 'Superman', 'Batman' ), 
        ( 'Superman', 'Ironman' ), 
        ( 'Superman', 'Wolverine' ), 
        ( 'Superman', 'Wolverine' ), 
        ( 'Batman', 'Wolverine' );

with pre as (
  SELECT
    [HeroName],
    count([HeroName]) [locationcount],
    '(' + ltrim(str(count([HeroName]))) + ') ' + [HeroFriend] [Concat]
  FROM @Heroes
  Group By [HeroName],
           [HeroFriend]
)

Select DISTINCT
       [HeroName],
       STUFF((SELECT ',' + [Concat]
              FROM (
                    SELECT TOP 100 [HeroName],[Concat]
                    FROM pre
                    order by locationcount desc
              ) tblThis 
              WHERE ([HeroName] = Result.[HeroName]
                 )
              FOR XML PATH ('')),1,1,'') AS [Loc Counts]
From pre AS Result

由于某种原因，当我尝试进行连接和/或动态排序时，SQL 不太喜欢它。代码可以正常运行，但忽略了排序。但是事先将它们放入“with”块中解决了它。

【讨论】：

【解决方案2】：

类似这样的：

with pre AS
(
  SELECT 
    [Client ID],
    ltrim(str([Store Num])) as [Store Num]
    count([Store Num]) as locationcount
  FROM @tbl_coreData
  Group By [Store Num], [Client ID]
)
Select [Client ID],
       STUFF((SELECT ', ' + [Location (counts)]
              FROM (
                SELECT [Client ID], [Store Num] + ' (' + locationcount + ') ' [Location (counts)]
                FROM pre
                Order By locationcount desc
              ) tblThis 
              WHERE ([Store Num] = Result.[Store Num] AND [Client ID] = Result.[Client ID]) 
              FOR XML PATH ('')),1,1,'') AS BATCHNOLIST
From pre AS Result
order by locationcount desc

【讨论】：

这很有帮助，因为我从未使用过“WITH”子句。我可以通过将 FROM pre Group By locationcount, [Store Num], [Client ID] 添加到子查询中来对值进行排序，但是默认情况下这是按升序排列的。 Order By 语句在子查询中不起作用，尽管它们也不会引发错误。
@n8 在子查询（pre）中应该没有理由按顺序排序。任何时候你使用它，你都会做一个选择，并可以在该选择中添加一个 order by 子句。我不确定您在哪里看不到正确的顺序。
我不明白为什么我对您的答案的编辑被拒绝了，我做了两个小修改以便它可以运行。您的子查询中有一个 order by，也许您的意思是排除它而不是让它工作？至于我上面的评论，我提到使用的 order by 最终没有起作用，看来我的初步结果是我的测试对象集的侥幸。所以这个问题仍然悬而未决。
@n8 子查询中的 orderby 应该改变逗号分隔结果的顺序。最后的 order by 应该改变行的顺序。