Oracle 替换0个或多个html标记中的文本

我有一个包含varchar2字符串的列的表，如下所示

My dog chases my cat
<p>My dog ate my other cat</p>
<p><div id="abcd">My cat ate my hamster</div><p>
<p><b><div id="abcd">My hamster sleeps all the time</div></b></p>

我需要对html标记中的文本执行SUBSTR

我认为编号组是一种方式，但我无法将结束标记放入自己的组中。这是我拥有的sql：

WITH sentences AS 
(
SELECT 1 as nr, 'My dog chases my cat' AS ln FROM DUAL
UNION
SELECT 2, '<p>My dog ate my other cat</p>' FROM DUAL
UNION
SELECT 3,'<p><x><div id="abcd">My cat ate my hamster</div></x></p>' FROM DUAL
UNION
SELECT 4,'<p><b><div id="abcd">My hamster sleeps all the time</div></b></p>' FROM DUAL
)
SELECT nr, regexp_replace(ln, '^((<[^>]+>)+)(.*)((<[^>]+>)+)$', 'group 1:\1,group 2:\2,group 3:\3,group 4:\4', 1, 1, 'n')
FROM sentences order by nr;

RETURNS

1   My dog chases my cat
2   group 1:<p>,group 2:<p>,group 3:My dog ate my other cat,group 4:</p>
3   group 1:<p><x><div id="abcd">,group 2:<div id="abcd">,group 3:My cat ate my hamster</div></x>,group 4:</p>
4   group 1:<p><b><div id="abcd">,group 2:<div id="abcd">,group 3:My hamster sleeps all the time</div></b>,group 4:</p>

---

第4组仅包含一个结束标记，其他结束标记位于第3组中，如第3行和第4行所示。我需要什么样的模式才能在他们自己的编号组中获得所有结束标记？

说过不要这样做，实际上对于这些特定值，您只剩下一个字符了：

WITH sentences AS 
(
SELECT 1 as nr, 'My dog chases my cat' AS ln FROM DUAL
UNION
SELECT 2, '<p>My dog ate my other cat</p>' FROM DUAL
UNION
SELECT 3,'<p><x><div id="abcd">My cat ate my hamster</div></x></p>' FROM DUAL
UNION
SELECT 4,'<p><b><div id="abcd">My hamster sleeps all the time</div></b></p>' FROM DUAL
)
SELECT nr,
  regexp_replace(ln, '^((<[^>]+>)+)(.*?)((<[^>]+>)+)$', 'group 1:\1,group 2:\2,group 3:\3,group 4:\4', 1, 1, 'n') as str
--------------------------------------^
FROM sentences order by nr;

或者只是那群人：

SELECT nr, regexp_replace(ln, '^((<[^>]+>)+)(.*?)((<[^>]+>)+)$', '\3', 1, 1, 'n') as str
FROM sentences order by nr;

NR STR                           
-- ------------------------------
 1 My dog chases my cat          
 2 My dog ate my other cat       
 3 My cat ate my hamster         
 4 My hamster sleeps all the time

---

说了不要这样做，对于这些特定的值，实际上只输出一个字符：

WITH sentences AS 
(
SELECT 1 as nr, 'My dog chases my cat' AS ln FROM DUAL
UNION
SELECT 2, '<p>My dog ate my other cat</p>' FROM DUAL
UNION
SELECT 3,'<p><x><div id="abcd">My cat ate my hamster</div></x></p>' FROM DUAL
UNION
SELECT 4,'<p><b><div id="abcd">My hamster sleeps all the time</div></b></p>' FROM DUAL
)
SELECT nr,
  regexp_replace(ln, '^((<[^>]+>)+)(.*?)((<[^>]+>)+)$', 'group 1:\1,group 2:\2,group 3:\3,group 4:\4', 1, 1, 'n') as str
--------------------------------------^
FROM sentences order by nr;

或者只是那群人：

SELECT nr, regexp_replace(ln, '^((<[^>]+>)+)(.*?)((<[^>]+>)+)$', '\3', 1, 1, 'n') as str
FROM sentences order by nr;

NR STR                           
-- ------------------------------
 1 My dog chases my cat          
 2 My dog ate my other cat       
 3 My cat ate my hamster         
 4 My hamster sleeps all the time

您可能能够使用这些特定字符串获得一些东西，但总的来说，这不是一个好主意。您可能能够使用这些特定字符串获得一些东西，但总的来说，这不是一个好主意。