WITH sample_data AS (SELECT 1 ID, 'a' val FROM dual UNION ALL
                     SELECT 1 ID, 'b' val FROM dual UNION ALL
                     SELECT 1 ID, 'c' val FROM dual UNION ALL
                     SELECT 2 ID, 'd' val FROM dual UNION ALL
                     SELECT 2 ID, 'e' val FROM dual UNION ALL
                     SELECT 2 ID, 'f' val FROM dual UNION ALL
                     SELECT 2 ID, 'g' val FROM dual UNION ALL
                     SELECT 3 ID, 'h' val FROM dual UNION ALL
                     SELECT 3 ID, 'i' val FROM dual UNION ALL
                     SELECT 3 ID, 'h' val FROM dual UNION ALL
                     SELECT 3 ID, 'j' val FROM dual UNION ALL
                     SELECT 3 ID, 'k' val FROM dual UNION ALL
                     SELECT 4 ID, 'l' val FROM dual UNION ALL
                     SELECT 4 ID, 'm' val FROM dual UNION ALL
                     SELECT 5 ID, 'n' val FROM dual)
select  substr(xmlcast(xmlagg(xmlelement(e,',' || val) order by val) as clob),2) vals
  from  sample_data
  group by id
/

VALS
--------------------------------------------------------------------------------
a,b,c
d,e,f,g
h,h,i,j,k
l,m
n

SQL>

with t as (
           select  id,
                   val,
                   sum(lengthb(val || ',')) over(partition by id order by val rows between unbounded preceding and current row) - 1 lb
             from  sample_data
          )
select  listagg(val,',') within group(order by val) limit_by_val
  from  t
  where lb <= 4000
  group by id
/