Oracle 字符串去重复

/*去重复*/
Select xmid,
regexp_replace( listagg(name,',') WITHin GROUP (ORDER BY xmid) ,'([^,]+)(,\1)*(,|$)','\1\3')as name
from t_tablename
Where yxzt=1 
--And xmid=4013903
Group By xmid

 select listagg(to_char(username),',') within group(order by username)name from Sys_User_Employee Where Rownum <3;

 

 

Select xmid, wm_concat(Distinct dbbm)  dbbm FROM (
select a.xmid,
level lev,
regexp_substr(a.dbbm,'[^,]+',1,level) as dbbm 
From (
Select xmid,(dbbm||','||phbm) dbbm FROM (
Select xmid,
regexp_replace( listagg(dbbm,',') WITHin GROUP  (ORDER BY xmid) ,'([^,]+)(,\1)*(,|$)','\1\3')as dbbm 
,(Select XMPHBM FROM T_PROJ_JBXX Where sysid=t.xmid) phbm
from t_proj_dbrw  t
Where yxzt=1 And cjrid='system' 
--And xmid=4013903
And xmid In(
4013923,
4013846,
4013920,
4013921,
4013844,
4013900,
4013901,
4013924
)
Group By xmid
)
)
 a 
Where a.dbbm Is Not Null
connect by prior a.xmid = a.xmid
and prior dbms_random.value is not null
and level <= regexp_count(a.dbbm,'[^,]+')
)

Group By xmid