Question

我有一个表，该表存储对象进度的连续周期的开始日期和结束日期。这是为了跟踪每个对象何时进入其进度的下一步

例如：

ObjectNum | Stage | StartDate | EndDate
1         |1      |2018-01-01 | 2018-02-24
1         |2      |2018-02-25 |
2         |1      |2018-01-01 |

SQL中是否有一种方法可以搜索此表并找到任何对象，其中一个阶段的结束日期与下一个连续阶段的开始之间的差超过1天？

亲切的问候

马特

Answer 1

您可以在ObjectNum和Stage上将表与其自身连接，然后进行测试：

SELECT
    t1.ObjectNum,
    t1.Stage,
    t1.StartDate,
    t1.EndDate,
    DATEDIFF("D", t1.EndDate, t2.StartDate) as StageGapInDays
FROM table t1
        INNER JOIN table t2 
            ON t1.ObjectNum = t2.ObjectNum
                AND t1.Stage + 1 = t2.Stage
Where StageGapInDays > 1

您还可以在子查询中使用LEAD()窗口函数，因此只需要扫描一次表并避免联接：

SELECT *
FROM
    (
        SELECT
            ObjectNum,
            Stage,
            StartDate,
            EndDate,
            DATEDIFF(day, EndDate, LEAD(StartDate) OVER (PARTITION BY ObjectNum ORDER BY Stage)) as StageGapInDays
        FROM table      
    ) subquery
Where StageGapInDays > 1

第二个将捕获Stage 1，否 Stage 2，然后是Stage 3和{{ {1}}大于一天。第一个查询无法解决这种情况。

Answer 2

针对对此感兴趣的人进行更新。考虑到Lead（）在SQL2008中不起作用，这是我采用的解决方案

Select ObjectNum from Object m join
       (SELECT ObjectNum, StartDate, EndDate, row_number() over (PARTITION by ObjectNum order by StartDate asc)as Row from STATHIS) s1
       on s1.ObjectNum = m.ObjectNum
       join
       (SELECT ObjectNum, StartDate, EndDate, row_number() over (PARTITION by ObjectNum order by StartDate asc)as Row from STATHIS) s2
       on s1.Row +1 = s2.Row and s1.ObjectNum = s2.ObjectNum
where DATEDIFF(DAY,s1.EndDate,s2.StartDate) > 1

检查单独的SQL行上的日期差异

2 个答案: