SQL查询仅返回5条记录而不是6行

Question

我试图获取用双哈希（##）分隔的单列中存在的数据。 根据下面提到的查询，我只能提取5条记录，而不是6行。

我可以认为connectby表达式存在一些问题。 任何帮助是极大的赞赏。

数据

Line1## Line2## Line3 ## Line4 ## Line5  ## Line6  ##

查询用于获取单个记录中的记录，这些记录用双井号##分隔

复制方案：

create table temp (errormessage varchar2(300))

insert into  temp errormessage values('Line1## Line2## Line3 ## Line4 ## Line5  ## Line6  ##')



WITH sample_data
     AS ( SELECT errormessage AS Error_Message
          FROM   TEMP )
SELECT Regexp_substr( error_message, ',?([^#]*)', 1, LEVEL, 'i', 1 ) AS Error_Message
FROM   sample_data
WHERE  Length( Regexp_substr( error_message, ',?([^#]*)', 1, LEVEL, 'i', 1 ) ) != 0
CONNECT BY ( Regexp_count(error_message, '#') + 1 >= LEVEL AND
             PRIOR dbms_random.value IS NOT NULL )
ORDER  BY LEVEL 

错误消息是具有要定界信息的列。 现在，很容易将问题复制到任何数据库中。

Answer 1

也许您正在追求类似：

WITH sample_data AS (SELECT 1 stagging_id,
                            'A' status,
                            'Line1## Line2## Line3 ## Line4 ## Line5  ## Line6  ##' error_message
                     FROM   dual UNION ALL
                     SELECT 2 stagging_id,
                            'B' status,
                            'Line1## Line2## Line3 ## Line4 ## Line5  ## Line6  ##Line7  ##' error_message
                     FROM   dual)
SELECT stagging_id,
       status,
       regexp_substr(error_message, '[^#]+', 1, LEVEL) err_msg
FROM   sample_data
CONNECT BY PRIOR stagging_id = stagging_id
           AND PRIOR sys_guid() IS NOT NULL
           AND regexp_substr(error_message, '[^#]+', 1, LEVEL) IS NOT NULL;

STAGGING_ID STATUS ERR_MSG
----------- ------ --------------------------------------------------------------
          1 A      Line1
          1 A       Line2
          1 A       Line3
          1 A       Line4
          1 A       Line5
          1 A       Line6
          2 B      Line1
          2 B       Line2
          2 B       Line3
          2 B       Line4
          2 B       Line5
          2 B       Line6
          2 B      Line7

现有代码的问题是regexp_substr中的* ，加上您正在计算单个#而分隔符为##的事实。

您可以这样修改查询：

WITH sample_data AS (SELECT 1 stagging_id,
                            'A' status,
                            'Line1## Line2## Line3 ## Line4 ## Line5  ## Line6  ##' error_message
                     FROM   dual UNION ALL
                     SELECT 2 stagging_id,
                            'B' status,
                            'Line1## Line2## Line3 ## Line4 ## Line5  ## Line6  ##Line7  ##' error_message
                     FROM   dual)
SELECT Regexp_substr( error_message, ',?([^#]+)', 1, LEVEL, 'i', 1 ) AS Error_Message
FROM   sample_dataCONNECT BY ( Regexp_count(error_message, '##') >= LEVEL AND
             PRIOR stagging_id = stagging_id AND
             PRIOR dbms_random.value IS NOT NULL )
ORDER  BY stagging_id, LEVEL;

ERROR_MESSAGE
--------------------------------------------------------------
Line1
 Line2
 Line3
 Line4
 Line5
 Line6
Line1
 Line2
 Line3
 Line4
 Line5
 Line6
Line7

请注意，我如何将regexp_substr中的* s更改为+ s，将regexp_count中的'#'更改为'##' 。