请问这个sql语句还有没有更优化的写法

zkyll

要取出表中某个字段的前n个最大的值，
设表名tab1,字段为f1
用 select * from tab1 where rownum<n order by f1 desc;
不行，因为这样取出来的是前n个物理记录，再在这n个值中进行排序。

用 select * from (select * from tab1 order by t1 desc) where rownum<n;
可以。

或者，建一个视图，create vtab1 as select * from tab1 order by t1 desc; 然后 select * from vtab1 where rownum<n;
也可以，就是感觉不够优化，表大了以后，特慢，不知各位大侠有没有更好的办法，请不吝赐教。

sshd

the 2nd SQL involves an inline view. it's basically the same as the 3rd SQL

You can try to build an index on the t1 column to spped up the SQL. Note the closer of ratio of the number of distinct t1 and the total number of t1 to 1, the better