更新了 Question-Oracle case 语句以提取与单行中的子字符串关联的所有事件的值

Updated Question-Oracle case statement to extract value for all occurance associated with a substring in a single line

我想编写一个 case 语句,它可以从名为 details 的列中提取特定子字符串的值,该列多次出现 [address] 是否可以将 REGEX 与 case 一起使用?

列中的示例数据:

[address]:kattengat 28
[address]:1012 SZ
[address]: Amsterdam

The below SQL only outputs:

kattengat 28

Select case when to_number(instr(details),'[address')>0 then substr(details,REGEXP_INSTR(details,'address',1,1)+8,instr(substr(details,REGEXP_INSTR(details,'address',1,1)+8,'[')-1)) else '' end from table_name;

Expected output is : 
kattengat 28 1012 SZ Amsterdam

创建 table 语句:

Create table test (id number(10), details clob);

插入语句:

insert into test (id, details) values (1,to_clob ('[ADDRESS    ] kattengat 28
             [NAME       ] ALEX
             [PHONE      ] 65438
             [ADDRESS    ] 1012 SZ
             [DOB        ] 1st Jan 1998
             [ADDRESS    ] Amsterdam')):

请注意,我不想连接和添加语句,而是寻找一种解决方案,该解决方案可以根据单个子字符串出现的次数提取与子字符串 [address] 关联的值行

这是一种选择:

SQL> with test (col) as
  2    (select '[address]:kattengat 28
  3  [address]:1012 SZ
  4  [address]: Amsterdam' from dual
  5    )
  6  select trim(replace(regexp_substr(replace(col, chr(10), '#'), '[^#]+', 1, column_value), '[address]:', '')) result
  7  from test cross join
  8    table(cast(multiset(select level from dual
  9                        connect by level <= regexp_count(col, ':')
 10                       ) as sys.odcinumberlist));

RESULT
--------------------------------------------------------------------------------
kattengat 28
1012 SZ
Amsterdam

SQL>

它有什么作用?

  • 第 1 - 5 行 - 样本数据
  • 第 6 行:
    • regexp_substr 部分代码负责将源列值拆分为单独的行
    • 它不影响原始值,而是影响换行符 (chr(10)) 被 # 替换(第二个 replace)的值,并且该字符用作regexp_substr
    • 的分隔符
  • 第一个 replace 从源中删除 [address]:
  • trim 删除 leading/trailing 个空字符串(如“Amsterdam”前面的字符串
  • 第 8 - 10 行用于从最终结果中删除重复值(如果源 table 不只包含一行,如我的示例所示)。如果确实如此,那么代码可以稍微简化。

使用您稍后发布的示例数据(顺便说一句,您确定 [NAME] 等前面有空格吗?我想没有!):

SQL> select * from test;

        ID DETAILS
---------- --------------------------------------------------
         1 [ADDRESS    ] kattengat 28
           [NAME       ] ALEX
           [PHONE      ] 65438
           [ADDRESS    ] 1012 SZ
           [DOB        ] 1st Jan 1998
           [ADDRESS    ] Amsterdam

我之前贴的代码,稍作修改,因为之前address是小写,方括号内没有空格,有冒号):

SQL> with temp as
  2    (select trim(replace(regexp_substr(replace(details, chr(10), '#'), '[^#]+', 1, column_value), '[ADDRESS    ]', '')) result
  3     from test cross join
  4     table(cast(multiset(select level from dual
  5                         connect by level <= regexp_count(details, '\[')
  6                        ) as sys.odcinumberlist))
  7    )
  8  select *
  9  from temp
 10  where instr(result, '[') = 0;

RESULT
--------------------------------------------------------------------------------
kattengat 28
1012 SZ
Amsterdam

SQL>

如果您想在一行中获得结果,您可以聚合该查询返回的值作为

SQL> with temp as
  2    (select trim(replace(regexp_substr(replace(details, chr(10), '#'), '[^#]+', 1, column_value), '[ADDRESS    ]', '')) result,
  3            column_value cv
  4     from test cross join
  5     table(cast(multiset(select level from dual
  6                         connect by level <= regexp_count(details, '\[')
  7                        ) as sys.odcinumberlist))
  8    )
  9  select listagg (result, ', ') within group (order by cv) final_result
 10  from temp
 11  where instr(result, '[') = 0;

FINAL_RESULT
--------------------------------------------------------------------------------
kattengat 28, 1012 SZ, Amsterdam

SQL>