Oracle listagg去重distinct的三种方式总结

select t.department_name depname,
       t.department_key,
       listagg(t.class_key, ',') within group(order by t.class_key) as class_keys
  from V_YDXG_TEACHER_KNSRDGL t
 where 1 = 1
 group by t.department_key, t.department_name

--第一种方法： 使用wm_concat() + distinct去重聚合
select t.department_name depname,
       t.department_key,
       wm_concat(distinct t.class_key) as class_keys
  from V_YDXG_TEACHER_KNSRDGL t
 where 1 = 1
 group by t.department_key, t.department_name

--第二种方法：使用正则替换方式去重(仅适用于oracle字符串大小比较小的情况)
select t.department_name depname,
       t.department_key,
       regexp_replace(listagg(t.class_key, ',') within
                      group(order by t.class_key),
                      '([^,]+)(,\1)*(,|$)',
                      '\1\3') as class_keys
  from V_YDXG_TEACHER_KNSRDGL t
 group by t.department_key, t.department_name;

--第三种方法：先去重，再聚合
select t.department_name depname,
       t.department_key,
       listagg(t.class_key, ',') within group(order by t.class_key) as class_keys
  from (select distinct s.class_key, s.department_key, s.department_name
          from V_YDXG_TEACHER_KNSRDGL s) t
 group by t.department_key, t.department_name
 
--或者
select s.department_key,
       s.department_name,
       listagg(s.class_key, ',') within group(order by s.class_key) as class_keys
  from (select t.department_key,
               t.department_name,
               t.class_key,
               row_number() over(partition by t.department_key, t.department_name, t.class_key order by t.department_key, t.department_name) as rn
          from V_YDXG_TEACHER_KNSRDGL t
         order by t.department_key, t.department_name, t.class_key) s
 where rn = 1
 group by s.department_key, s.department_name;