这里有一个选择;读取代码中的注解：
样本数据：

SQL> with scv_client_details  (match_details) as
  2    (select 'Matched with multiple QSIDs, updating with QSID: 6030066992. Possible QSIDs: 6030066992,1777,'
  3     from dual
  4     union all
  5     select 'Matched with multiple QSIDs. Yet another example; Possible QSIDs: 12345,6789'
  6     from dual
  7    ),

字符集
查询从这里开始：

8  temp as
  9    -- return strings that follow "Possible QSIDs:"
 10    (select trim(replace(substr(match_Details, instr(match_details, 'Possible QSIDs:')),
 11                    'Possible QSIDs:',
 12                    null
 13                   )) val
 14     from scv_client_details
 15    ),
 16  temp2 as
 17    -- split previously fetched comma-separated strings to rows. Avoid duplicate rows!
 18    (select regexp_substr(val, '[^,]+', 1, column_value) val
 19     from temp cross join
 20       table(cast(multiset(select level from dual
 21                           connect by level <= regexp_count(val, ',') + 1
 22                          ) as sys.odcinumberlist))
 23    )
 24  -- finally, result is numberic part of previously fetched VAL
 25  select regexp_substr(val, '\d+') result
 26  from temp2
 27  where regexp_like(val, '\d+');

型
测试结果：

RESULT
--------------------------------------------------------------------------------
6030066992
1777
12345
6789

SQL>

型