Question

    CustomerKey Person_Number Start_Date    End_Date
        1       250644    2017-01-01    4712-12-12
        2       250644    2017-01-10    4712-12-12
        3       250644    2017-01-15    4712-12-12

我需要SQL查询，其中下一个Start_Date-1和最后一个结束日期要更新的结束日期应为4712-12-12。

   CustomerKey  Person_Number Start_Date    End_Date
        1       250644      2017-01-01    2017-01-09
        2       250644      2017-01-10    2017-01-14
        3       250644      2017-01-15    4712-12-12

Answer 1

一种方法是使用lag()。这需要在update中进行自我联接：

update t
    set end_date = next_start_date - interval '1 day'
    from (select t.*,
                 lead(start_date) over (partition by person_number order by start_date) as next_start_date
          from t
         ) tt
    where tt.customerkey = t.customerkey and
          tt.next_start_date is not null;

尽管以上基本上适用于SQL Server（日期算术除外），但该问题最初被标记为Postgres。更好的SQL Server解决方案是：

update tt
    set end_date = dateadd(day, -1, next_start_date)
    from (select t.*,
                 lead(start_date) over (partition by person_number order by start_date) as next_start_date
          from t
         ) tt
    where tt.next_start_date is not null;

在SQL Server中不需要自联接。

Answer 2

根据当前值更新上一行：

update t1
set t1.End_Date = dateadd(day, -1, t2.Start_Date)
from tbl t1
inner join tbl t2 on t1.CustomerKey = (t2.CustomerKey - 1)

根据当前值更新上一行

2 个答案: