路径

Question

在sql server 2005中，我有一个描述路径的父表，以及一个包含每条路径上每个点的坐标的子表，如下所示：

路径

int path_id, string path_name

点

int path_id, float x, float y, int n

路径上的第一个点是n = 1，然后添加的每个新点都会递增n。

我需要做的是，删除此返回的点：

select [path_id]
from Points
where n=1

即。每个路径中的第一个点，然后还删除所有的点，其中n是由此返回的值：

select [path_id], max(n)
from Points
group by [path_id]

即。每个路径的最后一个添加点。

对于第一部分，我可以这样做：

delete from Points
where n=1

但我该如何处理第二部分？

Answer 1

DELETE FROM Points
FROM points
INNER JOIN     
(SELECT point_id, path_id, n
 FROM (SELECT ROW_NUMBER() OVER(PARTITION BY path_id ORDER BY n DESC) AS RowNumber, 
              point_id, path_id, n
       FROM points ) t
  WHERE n = 1 or RowNumber = 1 
) t on t.point_id = points.point_id and t.path_id = points.path_id and t.n = points.n

Answer 2

第二部分尝试这个：

DELETE from points where path_id in
(
   Select path_id from (
   select [path_id], max(n)
   from Points
   group by [path_id]
   ) t
)

Answer 3

我认为这个答案不适用于SQL Server，而是说明如何使用IN运算符支持元组的数据库：

DELETE FROM Points
WHERE (path_id, n) IN
    (SELECT path_id, max(n)
     FROM Points
     GROUP BY path_id)

Answer 4

这是一种可以解决问题的方法：

create table dummy as (
  select path_id, max(n)
  from points
  group by path_id
);

delete from points where n in 
  select n from dummy where points.path_id = dummy.path_id
);

drop table dummy;

如何将选择变为删除？

路径

点

4 个答案: