SQL 添加#前缀

Question

我有一个看起来像这样的选择查询

类型原因影响
增强数据传输化妆品
增强数据传输用户参与度
增强前端化妆品
增强前端用户参与度
问题数据传输化妆品
问题前端用户参与度
发布数据传输化妆品
发布前端用户参与度

我需要选择查询变成这样

类型原因影响
1.Enhancement 1.Data Transfer 11.Cosmetic
1.Enhancement 2.Data Transfer 12.User Engagement
1.增强 3.前端 13.化妆品
1.增强 4.前端 14.用户参与
2.问题 1.数据传输 21.化妆品
2.问题 2.前端 22.用户参与
3.问题 1.数据传输 31.化妆品
3.问题 2.前端 32.用户参与度

基本上第一列具有相同的前缀 # 每个唯一值，第二列的增量为 # 并重置每个第一列的值已更改。第三列只是第一列与第二列的组合。基于字母升序的值增加。

我认为这与行号和分区有关

Answer 1

你在寻找这样的东西吗？

select (dense_rank() over (order by type) || '.' || type) as ntype,
       (row_number() over (partition by type order by cause, impact)) || '.' || cause as ncause,
       (dense_rank() over (order by type) || row_number() over (partition by type order by cause, impact)) || '.' ||  impact) as nimpact
from t
order by ntype, ncause, nimpact;

Answer 2

这是我的答案，还有更简单的吗？

Select t1.Type,t1.Cause,Cast(t2.rnType as VARCHAR(9))+Cast(t3.rnCause as VARCHAR(9))+'.'+t1.Impact,t1.Tickets FROM
(SELECT [Type] as Type,(Case WHEN [Cause]='' THEN 'null' ELSE [Cause] END) as Cause,(Case WHEN [Impact]='' THEN 'null' ELSE [Impact] END) as Impact,count([Id]) as Tickets FROM {SupportTicket}
GROUP BY [Type],(Case WHEN [Cause]='' THEN 'null' ELSE [Cause] END),(Case WHEN [Impact]='' THEN 'null' ELSE [Impact] END)) t1
左连接
(SELECT [Type] as Type,dense_rank() over (ORDER By [Type]) as rnType FROM {SupportTicket}
分组依据 [类型]) t2
ON t1.Type=t2.Type
左连接
(SELECT [Type] as Type,[Cause] as Cause,dense_rank() over (PARTITION BY [Type] ORDER By [Cause]) as rnCause FROM {SupportTicket}
GROUP BY [类型],[原因]) t3
ON t1.Type=t3.Type 和 t1.Cause=t3.Cause

Answer 3

也许像

SELECT 
  concat(typeInc, '.', Type) as Type,
  concat(causeInc, '.', Cause) as Cause,
  concat(typeInc, causeInc, '.', Impact) as Impact
FROM (
  SELECT 
    TypeInc = DENSE_RANK() OVER (ORDER BY Type),
    CauseInc = ROW_NUMBER() OVER (PARTITION BY Type ORDER BY Cause),
    *
  FROM tbl
  ) withInc
 ORDER by Type, Cause, Impact;

http://sqlfiddle.com/#!18/bbcd95/27/0的演示

---

SQL 2008 兼容版本

SELECT 
  typeInc + '.' + Type as Type,
  causeInc + '.' + Cause as Cause,
  typeInc + causeInc + '.' + Impact as Impact
FROM (
  SELECT 
    TypeInc = CAST(DENSE_RANK() OVER (ORDER BY Type) AS varchar(3)),
    CauseInc = CAST(ROW_NUMBER() OVER (PARTITION BY Type ORDER BY Cause) AS varchar(3)),
    *
  FROM tbl
  ) withInc
 ORDER by Type, Cause, Impact;

http://sqlfiddle.com/#!18/bbcd95/29/0的演示