使用正则表达式提取逗号后的值

Question

我正在使用 Oracle 19c 数据库

下面是我的字符串值

variable B1 varchar2(60)
exec :B1:='(199,''TEST121''),(156,''TEST''),(1561,''TEST99'')';

我希望输出为

|    ID    |     NAME       |
| -------- | -------------- |
|       199|   TEST121      |
|       156|   TEST         |
|      1561|   TEST99       |

select  regexp_substr(regexp_substr(:b1,'[^A-Z+0-9][0-9]+', 1,level),'[0-9]+') as id , regexp_substr(:b1,'[A-Z]+[0-9]', 1,level) as name from dual connect by regexp_substr(:b1,'[0-9]', 1,level) is not null;

此查询仅输出以数字结尾的字符串值。

Answer 1

@Littlefoot 给出的选项的详细解释

下面是多值字符串转换为列和行的测试用例。

    set lines 999 pages 999
    col ID for a20
    col NAME for a20
    
    variable B1 varchar2(60)
    exec :B1:='(99,''TABLE1''),(56,''INDEX1''),(199,''TABLE''),(156,''INDEX'')';
    variable B1 varchar2(100)
    exec :B1:='(''TABL234E1~'',99),(''I1NDEX1~'',5ABC6),(''TAB1LE'',4ABC0),(''IND11EX'',6ACDE0)';
    
    WITH
       test (col)
       AS
          (SELECT :b1 FROM DUAL)
    SELECT SUBSTR (str, 1, INSTR (str, ',') - 1) id,
           SUBSTR (str, INSTR (str, ',') + 1) name
      FROM (    SELECT REGEXP_SUBSTR (
                          REPLACE (
                             REPLACE (
                                REPLACE (REPLACE (col, '),(', '#'), CHR (39), ''),
                                '(',
                                ''),
                             ')',
                             ''),
                          '[^#]+',
                          1,
                          LEVEL) str
                  FROM test
            CONNECT BY LEVEL <= REGEXP_COUNT (REPLACE (col, '),(', '#'), '#') + 1);

解释：

    variable B1 varchar2(60)
    
    exec :B1:='(99,''TABLE1''),(56,''INDEX1''),(199,''TABLE''),(156,''INDEX'')';

步骤1

    WITH
      test (col)
      AS
         (SELECT :b1 FROM DUAL)
      select col from test;

此步骤有助于获取 oracle 行格式的基本字符串。

结果：

     (99,'TABLE1'),(56,'INDEX1'),(199,'TABLE'),(156,'INDEX')

第二步

    WITH
      test (col)
      AS
         (SELECT :b1 FROM DUAL)
      select REPLACE(col, '),(','#') from test;

此步骤有助于将 '),(' 替换为 '#'（以获得更简单的多值字符串分隔符）。

结果：

    (99,'TABLE1'#56,'INDEX1'#199,'TABLE'#156,'INDEX')

第三步

    WITH
      test (col)
      AS
         (SELECT :b1 FROM DUAL)
      select REPLACE(REPLACE(col, '),(','#'),CHR(39),'') from test;

此步骤有助于替换上一步中的单引号。

结果：

    (TABL234E1~,99#I1NDEX1~,5ABC6#TAB1LE,4ABC0#IND11EX,6ACDE0)

第四步

    WITH
      test (col)
      AS
         (SELECT :b1 FROM DUAL)
      select REPLACE(REPLACE(REPLACE(col, '),(','#'),CHR(39),''),'(','') from test;

此步骤有助于将 '(' 替换为 ''（从前面的步骤中删除前导括号）。

结果：

    99,TABLE1#56,INDEX1#199,TABLE#156,INDEX)

步骤5

    WITH
      test (col)
      AS
         (SELECT :b1 FROM DUAL)
      select REPLACE(REPLACE(REPLACE(REPLACE(col, '),(','#'),CHR(39),''),'(',''),')','') from test;

此步骤有助于将 ')' 替换为 ''（从前面的步骤中删除尾括号）。

结果：

   99,TABLE1#56,INDEX1#199,TABLE#156,INDEX

步骤6

    WITH
       test (col)
       AS
          (SELECT :b1 FROM DUAL)
    SELECT *
      FROM (    SELECT REGEXP_SUBSTR (
                          REPLACE (
                             REPLACE (
                                REPLACE (REPLACE (col, '),(', '#'), CHR (39), ''),
                                '(',
                                ''),
                             ')',
                             ''),
                          '[^#]+',
                          1,
                          LEVEL) str
                  FROM test
            CONNECT BY LEVEL <= REGEXP_COUNT (REPLACE (col, '),(', '#'), '#') + 1);

这一步有两个部分：

    a. REGEXP_SUBSTR(<value from previous step>,'[^#]+',1,LEVEL)
    '[^#]+' ==> searches and gets every character other than '#' 
    1==> starting position of the searches
    LEVEL==> LEVEL can be used in conjunction with CONNECT BY LEVEL clause. All the chunks of the source string can be displayed by using the LEVEL keyword as the match occurrence. 

    b. CONNECT BY LEVEL <= REGEXP_COUNT (REPLACE (col, '),(', '#'), '#') + 1
    REPLACE (col, '),(', '#') ==> This helps in replacing '),(' with '#' (to get a simpler separator for multivalued string).
    REGEXP_COUNT(<output from previous step>,'#')+1==> Count the number of '#' +1 in the source string. 
    Here, the CONNECT BY LEVEL clause generates the rows equal to the number of '#' +1 in the source string. 

结果：

    99,TABLE1
    56,INDEX1
    199,TABLE
    156,INDEX

步骤7

    WITH
       test (col)
       AS
          (SELECT :b1 FROM DUAL)
    SELECT SUBSTR (str, 1, INSTR (str, ',') - 1) id,
           SUBSTR (str, INSTR (str, ',') + 1) name
      FROM (    SELECT REGEXP_SUBSTR (
                          REPLACE (
                             REPLACE (
                                REPLACE (REPLACE (col, '),(', '#'), CHR (39), ''),
                                '(',
                                ''),
                             ')',
                             ''),
                          '[^#]+',
                          1,
                          LEVEL) str
                  FROM test
            CONNECT BY LEVEL <= REGEXP_COUNT (REPLACE (col, '),(', '#'), '#') + 1);
    

这一步有 2 个组成部分： a.SUBSTR (str, 1, INSTR (str, ',') - 1)==> str 是从前面的输出中提取的，从第一个位置开始，它将字符从 ',' greps 到 -1 位置。 b.SUBSTR (str, INSTR (str, ',') + 1) name==> str 是从前面的输出中提取的，从 ',' 开始，它将字符从 ',' 到 +1 位置。

结果：

    ID                   NAME
    -------------------- --------------------
    99                   TABLE1
    56                   INDEX1
    199                  TABLE
    156                  INDEX

Answer 2

这是一种选择：

SQL> WITH
  2     test (col)
  3     AS
  4        (SELECT '(199,''TEST121''),(156,''TEST''),(1561,''TEST99'')' FROM DUAL)
  5  SELECT SUBSTR (str, 1, INSTR (str, ',') - 1) id,
  6         SUBSTR (str, INSTR (str, ',') + 1) name
  7    FROM (    SELECT REGEXP_SUBSTR (
  8                        REPLACE (
  9                           REPLACE (
 10                              REPLACE (REPLACE (col, '),(', '#'), CHR (39), ''),
 11                              '(',
 12                              ''),
 13                           ')',
 14                           ''),
 15                        '[^#]+',
 16                        1,
 17                        LEVEL) str
 18                FROM test
 19          CONNECT BY LEVEL <= REGEXP_COUNT (REPLACE (col, '),(', '#'), '#') + 1);

ID                               NAME
-------------------------------- --------------------------------
199                              TEST121
156                              TEST
1561                             TEST99

SQL>

它有什么作用？

• 第 1 - 4 行 - 示例数据
• 第 8 - 14 行 - 将 ),( 替换为 #（以获得更简单的分隔符）；删除前括号和尾括号
• 第 7 - 19 行 - 将示例字符串拆分为行
• 第 5、6 行 - 从每一行中提取 ID 和 NAME