// cluster_by_rowid // clustering factor 가 낮을 때 사용하면 빠른 결과 도출 select /*+ ordered use_nl(b) rowid(b) */ b.* from (select /*+ index(emp emp_X1) no_merge */ rowid rid from emp where ename = 'SCOTT' order by rowid) a, emp b where b.rowid = a.rid ; -- after 12c select /*+ cluster_by_rowid(e) index(e emp_x1) */ * from emp e where ename = 'SCOTT' ;