Regex Oracle SQL：提取每行中所有匹配的字符串

假设我有一个带有两列ITEM和CONTENT的表MESSAGE

目前，我想用这个模式${variable}显示每个单词。预期的最终结果是

ITEM1 | ${username}, ${address}
ITEM2 | ${username}

要实现这一点，正确的Oracle SQL查询是什么？我尝试过类似的东西，但它只是列出了与匹配的正则表达式的内容

SELECT ITEM, REGEXP_SUBSTR(CONTENT, '^.*\$\{.*\}.*$', 1, 1) FROM MESSAGE;

---

我不太擅长正则表达式，看看这是否有帮助：

SQL> with message (item, content) as
  2    (select 'ITEM1', 'Dear ${username}, you have changed your address to ${address}' from dual union
  3     select 'ITEM2', 'Hi ${username}, thank you for attending this event.' from dual
  4    )
  5  select item,
  6    regexp_substr(content, '\$\{\w+}', 1, 1) the_first,
  7    regexp_substr(content, '\$\{\w+}', 1, 2) the_second
  8  from message;

ITEM  THE_FIRST            THE_SECOND
----- -------------------- --------------------
ITEM1 ${username}          ${address}
ITEM2 ${username}

SQL>

它有什么作用

\$在这里用于转义美元符号，因为它用作字符串结尾的锚点，所以-转义它 \{也一样，因为花括号表示出现的次数，所以-避开它 \w+代表整个世界 1,1和1,2：从第一个字符开始，取该表达式的第一个1,1或第二个1,2外观

---

将这两个表达式连接起来并用逗号分隔它们是很简单的。

我不太擅长正则表达式，看看这是否有帮助：

SQL> with message (item, content) as
  2    (select 'ITEM1', 'Dear ${username}, you have changed your address to ${address}' from dual union
  3     select 'ITEM2', 'Hi ${username}, thank you for attending this event.' from dual
  4    )
  5  select item,
  6    regexp_substr(content, '\$\{\w+}', 1, 1) the_first,
  7    regexp_substr(content, '\$\{\w+}', 1, 2) the_second
  8  from message;

ITEM  THE_FIRST            THE_SECOND
----- -------------------- --------------------
ITEM1 ${username}          ${address}
ITEM2 ${username}

SQL>

它有什么作用

\$在这里用于转义美元符号，因为它用作字符串结尾的锚点，所以-转义它 \{也一样，因为花括号表示出现的次数，所以-避开它 \w+代表整个世界 1,1和1,2：从第一个字符开始，取该表达式的第一个1,1或第二个1,2外观 连接这两个字符并用逗号分隔它们是很简单的。

您可以使用\$\{.+？\}-使用？使之成为{}之间一个或多个字符的非贪婪匹配 捕获该组

使用具有Previor的connect by循环和具有level的SYS_GUID来提取一行中所有可能的匹配项

LISTAGG进行连接

WITH message AS (
    SELECT
        'ITEM1' AS item,
        'Dear ${username}, you have changed your address to ${address}' AS content
    FROM
        dual
    UNION ALL
    SELECT
        'ITEM2',
        'Hi ${username}, thank you for attending this event.'
    FROM
        dual
) 
SELECT item,LISTAGG(content,',') WITHIN GROUP ( ORDER BY lvl )
FROM
(
SELECT
    item,
    regexp_substr(content,'(\$\{.+?\})',1,level,NULL,1) as content,
    level as lvl
  FROM
      message
CONNECT BY level <= regexp_count(content,'\$\{.+?\}')
           AND PRIOR item = item
           AND PRIOR sys_guid() IS NOT NULL
           ) GROUP BY item;

您可以使用\$\{.+？\}-使用？使之成为{}之间一个或多个字符的非贪婪匹配 捕获该组

使用具有Previor的connect by循环和具有level的SYS_GUID来提取一行中所有可能的匹配项

LISTAGG进行连接

WITH message AS (
    SELECT
        'ITEM1' AS item,
        'Dear ${username}, you have changed your address to ${address}' AS content
    FROM
        dual
    UNION ALL
    SELECT
        'ITEM2',
        'Hi ${username}, thank you for attending this event.'
    FROM
        dual
) 
SELECT item,LISTAGG(content,',') WITHIN GROUP ( ORDER BY lvl )
FROM
(
SELECT
    item,
    regexp_substr(content,'(\$\{.+?\})',1,level,NULL,1) as content,
    level as lvl
  FROM
      message
CONNECT BY level <= regexp_count(content,'\$\{.+?\}')
           AND PRIOR item = item
           AND PRIOR sys_guid() IS NOT NULL
           ) GROUP BY item;

---

这假设每个内容最多有两个变量。OP的问题不清楚该假设是否有效。这假设每个内容最多有两个变量。OP的问题不清楚该假设是否有效。