它必须是regex吗？substr以简单的方式处理它：

SQL> with station (city) as
  2    (select 'Zagreb' from dual union all
  3     select 'Athens' from dual union all
  4     select 'Rome'   from dual union all
  5     select 'Ottawa' from dual
  6    )
  7  select distinct city
  8      from station
  9      where upper(substr(city, 1, 1)) not in ('A', 'E', 'I', 'O', 'U')
 10        and upper(substr(city, -1))   not in ('A', 'E', 'I', 'O', 'U');

CITY
------
Zagreb

SQL>

如果必须使用regex，则

7  select distinct city
  8      from station
  9      where regexp_like(city, '^[^aeiou].*[^aeiou]$', 'i');

CITY
------
Zagreb

SQL>