Question

我有这个sql查询，它通过movie_name选择它在表中找到的重复项：

SQL：

SELECT movies.movie_name, movies.year FROM movies
INNER JOIN (SELECT movie_name FROM movies
GROUP BY movie_name HAVING count(movie_id) > 1) dup ON movies.movie_name = dup.movie_name
// want also to test for same year, not just movie_name i.e movies.year = dup.year

这可能吗？

Answer 1

SELECT movies.movie_name, movies.year FROM movies
INNER JOIN (SELECT movie_name, year FROM movies
GROUP BY movie_name,year HAVING count(movie_id) > 1) dup ON movies.movie_name = dup.movie_name
and movies.year = dup.year

似乎是一个合理的开始...

删除一个我认为你的意思是保留一个，不要忘记你可以有一个以上的重复

让我们说我们会保留第一个摆脱其余的，而最早的movie_id是第一个

所以

Select Min(Movie_id), Movie_Name, Year From Movies Group By Movie_Name, Year

会给你所有的东西

Select Movie_id,Movie_Name,Year From Movies m 
Left Join 
(Select Min(Movie_id), Movie_Name, Year From Movies Group By Movie_Name, Year) keep
On keep.movieid = m.movieid Where keep.Movie_Id is null

以上是电影中的所有记录，但不是我们想要保留的所有记录。

这样就可以为我们提供你想要摆脱的所有东西。因为克苏鲁的缘故，不要相信我或你自己！在删除之前进行备份！

Delete m From Movies m 
Left Join 
(Select Min(Movie_id), Movie_Name, Year From Movies Group By Movie_Name, Year) keep
On keep.movieid = m.movieid Where keep.Movie_Id is null

所以现在我们已经接受了我们证实的查询（你确实证明了它不是你！）而不是选择我们正在删除它们的违规记录。

不要忘记备份！

Answer 2

我非常喜欢Tony使用MIN的想法。但是，整个查询可以更简单：

DELETE FROM movies WHERE movie_id NOT IN 
(SELECT MIN(movie_id) FROM movies GROUP BY movie_name, year);

用两个条件查询内连接？

2 个答案: