Mysql随机查询调优方法

在mysql中查询10条不重复的数据，使用以下：

SELECT * FROM `tableName` ORDERBY RAND()  LIMIT 10

就可以了。但是真正测试一下才发现这样效率非常低。一个21万余条的库，查询10条数据，居然要10秒以上

搜索Google，网上基本上都是查询max(id) * rand()来随机获取数据。

SQL语句如下：

SELECT * FROM `fb_tb_goods` AS t1 JOIN (SELECT ROUND(RAND() * ((SELECT MAX(id) FROM `fb_tb_goods` WHERE status in (1,3) AND cj_level in (2,3) )-(SELECT MIN(id) FROM `fb_tb_goods` WHERE status in (1,3) AND cj_level in (2,3)))+(SELECT MIN(id) FROM `fb_tb_goods` WHERE status in (1,3) AND cj_level in (2,3))) AS aid) AS t2 WHERE t1.id >= t2.aid and t1.status in (1,3)ORDER BY t1.id LIMIT 16

我测试了一下，需要0.5秒，速度也不错，但是跟上面的语句还是有很大差距。总觉有什么地方不正常。

于是我把语句改写了一下。

SELECT * FROM `fb_tb_goods` AS t1 JOIN (SELECT ROUND(RAND() * ((SELECT MAX(id) FROM `fb_tb_goods`)-(SELECT MIN(id) FROM `fb_tb_goods`))+(SELECT MIN(id) FROM `fb_tb_goods`)) AS aid) AS t2 WHERE t1.id >= t2.aid and t1.status in (1,3) ORDER BY t1.id LIMIT 26;

这下，效率又提高了，查询时间只有0.01秒

最后，再把语句完善一下，加上MIN(id)的判断。我在最开始测试的时候，就是因为没有加上MIN(id)的判断，结果有一半的时间总是查询到表中的前面几行。

The End