数据库oracle/mysql多条重复数据，取最新的

MYSQL

1、分组取最新的，再用最新的字段连表，回查数据

 ①按分组字段取最新的，通过主键id回表查数据

select * from table_name
WHERE id in (
select max(id) id from table_name GROUP BY 分组字段)

②非主键字段

SELECT r.*
FROM records r
INNER JOIN (
    SELECT group_id, MAX(timestamp) as latest
    FROM records
    GROUP BY group_id
) as latest_records 
ON r.group_id = latest_records.group_id 
AND r.timestamp = latest_records.latest;

ORACLE

1、原理讲解-可直接看2

筛选出最新的

SELECT *  
FROM (  
    SELECT t.*,  
           ROW_NUMBER() OVER (PARTITION BY LOCALAUTHID ORDER BY LASTUPDATETIME DESC) AS rn  
    FROM USER_LOCALAUTH_STATE t  
) t  
WHERE t.rn = 1;

解释：

这个序号是基于[LOCALAUTHID]字段进行分区的（即每个[LOCALAUTHID]值都是一个新的分区），并且在每个分区内部，行是按照LASTUPDATETIME字段降序排列的

2、在重复数据取最新的

SELECT *  
FROM (  
    SELECT t.*,ROW_NUMBER() OVER (PARTITION BY LOCALAUTHID ORDER BY LASTUPDATETIME DESC) AS rn  
    FROM    
		(
			-- 筛选存在重复的数据，count(LOCALAUTHID) > 1
			select * from USER_LOCALAUTH_STATE
			where LOCALAUTHID	in (
				SELECT LOCALAUTHID
				FROM USER_LOCALAUTH_STATE WHERE STATE = 1 AND BUSINESSTYPE=1
				group by LOCALAUTHID having count(LOCALAUTHID) > 1)
		) t
	) rt  
WHERE rt.rn = 1;

3、若要删除不满足条件

在2的基础上，通过[LOCALAUTHID]查询，并删除ID不再2中的