如果您希望所有这些都作为 row 中的单个字符串，则无需使用正则表达式，您可以使用标准 REPLACE() :

If you want all of these as a single string in a row them there's no need to use regular expressions you can use a standard REPLACE():

SQL> select replace('2711393|2711441|1234567', '|', ', ') from dual;

REPLACE('2711393|2711441|
-------------------------
2711393, 2711441, 1234567

如果要将所有这些都放在一个列中，则需要使用 CONNECT BY ，因为我演示了

If you want all of these in a single column then you need to use CONNECT BY as I demonstrate here. Please note that this is highly inefficient.

SQL>  select regexp_substr('2711393|2711441|1234567', '[^|]+', 1, level)
  2     from dual
  3  connect by regexp_substr('2711393|2711441|1234567'
  4                           , '[^|]+', 1, level) is not null;

REGEXP_SUBSTR('2711393|2711441|1234567','[^|]+',1,LEVEL)
--------------------------------------------------------------------------

2711393
2711441
1234567

SQL>

如果要在不同的列中使用它们，则需要使用

If you want these in different columns you need to use PIVOT and you'll need to know how many you have. I'm assuming 3.

SQL> select *
  2    from (
  3   select regexp_substr('2711393|2711441|1234567', '[^|]+', 1, level) as a
  4        , level as lvl
  5     from dual
  6  connect by regexp_substr('2711393|2711441|1234567'
  7                           , '[^|]+', 1, level) is not null
  8          )
  9   pivot ( max(a)
 10          for lvl in (1,2,3)
 11          )
 12         ;

1          2          3
---------- ---------- ----------
2711393    2711441    1234567

SQL>

如您所见，这些都是完全可怕的，除了第一个，效率很低.您应该正确规范化数据库，以确保不必这样做.

As you can see these are all completely horrible and, save the first, highly inefficient. You should normalise your database correctly to ensure you don't have to do this.