SELECT 语句，带有附加逻辑，用于修改某些记录

Posted 2023-05-08

【中文标题】SELECT 语句，带有附加逻辑，用于修改某些记录

【英文标题】：SELECT Statement, with additional logic, in order to modify certain records

【发布时间】：2021-01-19 15:30:52

【问题描述】：

如何制定也可以修改Query 结果的SELECT 语句。

我有下表：

ID    Startdate     Enddate        Category

0,  "2021-05-05", "2021-05-20", "First Category" 
1,  "2021-05-06", "2021-05-10", "First Category"
2,  "2021-05-07", "2021-05-10", "First Category"
3,  "2021-05-08", "2021-05-15", "First Category"
4,  "2021-05-20", "2021-05-22", "Second Category"
5,  "2021-05-21", "2021-05-24", "Second Category"
6,  "2021-05-23", "2021-05-24", "First Category"

获取所有这些记录的 SQL 语句非常基本，但是我需要属于一个类别序列（0-3、4-5、6）的所有记录，Startdate 中的所有记录序列，必须设置为序列第一条记录的Startdate。

因此，预期结果如下：

ID    Startdate     Enddate        Category

0,  "2021-05-05", "2021-05-20", "First Category" 
1,  "2021-05-05", "2021-05-10", "First Category"  //changed Startdate
2,  "2021-05-05", "2021-05-10", "First Category"  //changed Startdate
3,  "2021-05-05", "2021-05-15", "First Category"  //changed Startdate
4,  "2021-05-20", "2021-05-22", "Second Category"
5,  "2021-05-20", "2021-05-24", "Second Category" //changed Startdate
6,  "2021-05-23", "2021-05-24", "First Category"

【参考方案1】：

这是一种间隙和孤岛问题，但您不想减少行数，只是为了存储信息。我建议运行max() 来确定“岛”的开始位置，累积总和来计算每行所在的“岛”。最后是您想要的窗口函数：

select t.*,
       min(startdate) over (partition by category, grp) as group_startdate
from (select t.*,
             sum(case when prev_enddate >= startdate then 0 else 1 end) over (partition by category order by startdate) as grp
      from (select t.*,
                   max(enddate) over (partition by category order by startdate rows between unbounded preceding and 1 preceding) as prev_enddate
            from t
           ) t
     ) t;

Here 是一个 dbfiddle。