NewID()用法

前天碰到一业务需求，难倒了团队成员，

表结构如下： CREATE TABLE [dbo].[Product](
[P_ID] [bigint] IDENTITY(1,1) NOT NULL,
[P_Name] [nvarchar](255) NULL,
[CategoryID1] [int] NULL,
[CategoryID2] [int] NULL,
[CategoryID3] [int] NULL,
[P_SingleIntro] [nvarchar](200) NULL,
[LoginID] [nvarchar](50) NULL,
CONSTRAINT [PK_Product] PRIMARY KEY CLUSTERED
(
[P_ID] ASC
)

需要随机列出表中每位用户(loginid)的一个产品，每次列出时随机值不重复。

于是考虑用newid() select max(P_ID)as P_ID,loginID from product
group by loginid order by NewID()

结果每次取到的P_ID都是相同的！　不符合需求

再修改如下： select P_ID,LoginID,P_Name,P_SingleIntro from product where P_ID in
(
select ( select top 1 p_id from product as b where b.loginid = c.loginid order by newid() ) as p_id
from ( select top 10000 a.loginID from product as a group by a.loginid order by NewID() ) as c
)

--假定取前10000个用户