这里有一个选择；阅读代码中的注解。我不想打字太多，所以要两次约会。
示例数据（您已经有了&不要键入它。您需要的代码从第10行开始：

SQL> with
  2  -- sample data
  3  test (datum, total, col) as
  4    (select date '2020-07-20', 100, '10,0,20,30,0' from dual union all
  5     select date '2020-07-20', 150, '15,3,40,30,2' from dual union all
  6     --
  7     select date '2020-07-19', 200, '50,6,50,30,8' from dual union all
  8     select date '2020-07-19', 300, '20,1,40,10,2' from dual
  9    ),

将csv值拆分为行。注意 RB 有助于我们求和匹配值的值

10  -- split comma-separated values into rows
 11  temp as
 12    (select
 13        datum,
 14        total,
 15        to_number(regexp_substr(col, '\d+', 1, column_value)) val,
 16        column_value rb
 17      from test cross join
 18        table(cast(multiset(select level from dual
 19                            connect by level <= regexp_count(col, ',') + 1
 20                           ) as sys.odcinumberlist))
 21    ),

计算摘要简单；没什么特别的。我们将保留 RB 在最后一步中需要的值：

22  -- compute summaries
 23  summary as
 24    (select datum,
 25       sum(total) total,
 26       sum(val) sumval,
 27       rb
 28     from temp
 29     group by datum, rb
 30    )

最后一步。使用 LISTAGG ，将逗号分隔的值聚合回，但这次将彼此相加：

31  -- final result
 32  select datum,
 33    total,
 34    listagg(sumval, ',') within group (order by rb) new_col
 35  from summary
 36  group by datum, total
 37  order by datum desc, total;

DATUM                    TOTAL NEW_COL
------------------- ---------- --------------------
20.07.2020 00:00:00        250 25,3,60,60,2
19.07.2020 00:00:00        500 70,7,90,40,10

SQL>