保留id最小的记录

查询重复数据
select * from TABLE_NAME where (movie_id,star_id) in (select movie_id,star_id from TABLE_NAME group by movie_id,star_id having count() > 1) and id not in (select min(id) from TABLE_NAME group by movie_id,star_id having count()>1)

删除重复数据
delete from TABLE_NAME where (movie_id,star_id) in (select movie_id,star_id from TABLE_NAME group by movie_id,star_id having count() > 1) and id not in (select min(id) from TABLE_NAME group by movie_id,star_id having count()>1)

sqlSELECT movie_id, star_id, COUNT(*) FROM xxx GROUP BY (movie_id + '-' + star_id) HAVING COUNT(*) > 1

我是有一个和你差不多的，但是我高并发，一分钟3000条数据入库。
我是先将数据放memcahe里面，
每个入库的先匹配，
如果没有，就入库，
如果有的，比较数据时间戳，再做处理。。。。

SELECT distinct movie_id, star_id FROM xxx

SELECT t.id, t.movie_id, t.star_id from xxx AS t LEFT JOIN xxx AS t1 on t.movie_id = t1.movie_id WHERE t.id != t1.id AND t.star_id = t1.star_id

把xxx换成你的表名，这里的id是主键id，如果你表里没有那就换成time，但是这样可能会造成time也相同的重复数据无法被找到。