你可以试试 split 以及 explode ```
SELECT DISTINCT * FROM
(
select explode( split('a,b,b,c,d,e,f' ,',')) c
) as t
ORDER BY c

你也可以用 stack ，它的工作速度比union快得多：

hive> select id from
    > (
    > select stack(5, --the number of elements
    > 10 ,
    > 20 ,
    > 30 ,
    > 40 ,
    > 50
    > ) as (id)
    > )s;
OK
10
20
30
40
50
Time taken: 4.88 seconds, Fetched: 5 row(s)

SELECT DISTINCT *
FROM (
    SELECT 12
    UNION SELECT 23
    UNION SELECT 34
    UNION SELECT 12
    UNION SELECT 23
) AS t1;

ps：另请参阅非常密切相关的问题：如何从sql server中的值列表中进行选择