with t as
(
    select '2011-08-22' as date_id, 'tag1' as biz_tag from dual
     union all
    select '2011-08-22' as date_id, 'tag2' as biz_tag from dual
     union all
    select '2011-08-22' as date_id, 'tag3' as biz_tag from dual
)
select tab1.date_id,
       replace(wmsys.wm_concat(case when tab1.biz_tag = 'tag1' then tab1.biz_tag else '' end), ',' , '') as biz_tag1,
       replace(wmsys.wm_concat(case when tab1.biz_tag = 'tag2' then tab1.biz_tag else '' end), ',' , '') as biz_tag2,
       replace(wmsys.wm_concat(case when tab1.biz_tag = 'tag3' then tab1.biz_tag else '' end), ',' , '') as biz_tag3
  from (select date_id, biz_tag from t group by date_id, biz_tag) tab1
 group by tab1.date_id;


DATE_ID    BIZ_TAG1  BIZ_TAG2  BIZ_TAG3
---------- --------- --------- --------
2011-08-22 tag1      tag2      tag3

------解决方案--------------------
SQL code
----date_id biz_tag1 biz_tag2 biz_tag3
2011-08-22 tag1 tag2 tag3
select date_id,
sum(decode(biz_tag,tag1,null)) biz_tag1 ,
sum(decode(biz_tag,tag2,null)) biz_tag2,
sum(decode(biz_tag,tag3,null)) biz_tag3
from tb
group by date_id

------解决方案--------------------
3楼写的复杂了啊
SQL code

with t as
(
    select '2011-08-22' as date_id, 'tag1' as biz_tag from dual
     union all
    select '2011-08-22' as date_id, 'tag2' as biz_tag from dual
     union all
    select '2011-08-22' as date_id, 'tag3' as biz_tag from dual
)
SELECT date_id,
MAX(decode(biz_tag,'tag1','tag1','')) biz_tag1,
MAX(decode(biz_tag,'tag2','tag2','')) biz_tag2,
MAX(decode(biz_tag,'tag3','tag3','')) biz_tag3 FROM t
GROUP BY date_id