Question

我创建了以下代码，以便插入两个表Users和Roles：

- 插入Users表

declare @u int
set @u = 1000

while @u <= 1200
begin
   insert into Users(UserID, Username, [Password], Email)
   values(@u, 
         'Username'+CAST(@u as varchar(4)),
         'pass'+CAST(@u as varchar(4)),
         'Email'+CAST(@u as varchar(4)))

   set @u = @u + 1
end
--to insert into the Roles table
declare @a int
set @a = 1

while @a <= 100
begin
  insert into Roles (RoleID, [Role], [Description])
  values(@a, 'Admin', 'description’ + CAST(@a as varchar(4)))

  set @a = @a + 1
end

现在我想对一个名为'User_Roles'的表做同样的事情，它从Users表中得到两个外键，从Roles得到另外一个。我写了以下内容：

declare @b int
set @b = 1

while @b <= 300
begin
  insert into User_Roles(UserID, RoleID, [Description])
    select  
       UserID, RoleID, 'Description' + CAST(@b as varchar(4))
    from Users, Roles
    where UserID = CAST((RAND()*200+1000) as int)
      and RoleID = CAST((RAND()*99+1)as int)

  set @b = @b + 1
end

两个第一个代码将正确执行，但最后一个代码会导致错误，因为它会在主键中插入重复值。如何更改最后一部分并解决此问题？

Answer 1

修改INSERT语句以包含GROUP BY，以消除UserID和RoleID重复项。

 insert into User_Roles(UserID, RoleID,[Description])
     select  UserID, RoleID,'Description'+CAST(@b as varchar(4))
     from Users, Roles
     where  UserID = CAST((RAND()*200+1000) as int)
     and    RoleID = CAST((RAND()*99+1)as int)
     GROUP BY UserID, RoleID

Answer 2

每次循环时，都必须确保插入一个尚不存在的值。

未经测试，但这样的事情应该有效：

  insert into User_Roles(UserID, RoleID, [Description]) 
    select   
       u.UserID, r.RoleID, 'Description' + CAST(@b as varchar(4)) 
    from Users u, Roles r
    where u.UserID = CAST((RAND()*200+1000) as int) 
      and r.RoleID = CAST((RAND()*99+1)as int) 
      and NOT EXISTS(SELECT 1 from User_Roles ur WHERE ur.UserID = u.UserID AND ur.RoleId = r.RoleId)

帮助解决SQL Server分配中的问题

2 个答案: