Pivoting and string manipulation

Question

Hi Guys I would appreciate your help with this.
I have a table with a single column= A1, A2, B1, C1, C2 which I need to pivot and write the following rows to a second table again as a single column;
A1,B1,C1
A1,B1,C2
A2,B1,C1
A2,B1,C2
An example would be great. Thanks in advance.
Gaz

Answer 1

Hi,
Gaz wrote:
... Hope this covers all your questionsNo.
Frank Kulash wrote:
... Explain the business rules as clearly as you can. Why don't you want rows like the following in the second table:
C1, B1, A1
B1, B1, B1
A1, A2, B1or do you?
Do you know, in advance, that the strings all start with 'A', 'B' or 'C', or does the query have to figure out what the different possibilites are?I made several guesses.
INSERT INTO     target_tab (col1)
WITH  normalized_source_tab     AS
     SELECT     REGEXP_SUBSTR ( col1
                            , '[^ ,]+'
                     , 1
                     , LEVEL
                     )          AS txt
     ,     DENSE_RANK () OVER ( ORDER BY  REGEXP_SUBSTR ( col1
                                                                 , '[^ ,0-9]+'
                                              , 1
                                              , LEVEL
                             )          AS prefix_num
     FROM     source_tab
     CONNECT BY     LEVEL <= 1 + LENGTH (col1)
                             - LENGTH (REPLACE (col1, ','))
SELECT     SUBSTR ( SYS_CONNECT_BY_PATH (txt, ', ')
            , 3
FROM     normalized_source_tab
WHERE     CONNECT_BY_ISLEAF     = 1
START WITH     prefix_num     = 1
CONNECT BY     prefix_num     = 1 + PRIOR prefix_num     
;Among other things, I guessed that the comma-delimited items in sorce_tab.col1 always follow this pattern:
(1) Optional spaces, to be ignored
(2) Non-digits, which deterrmine the postion in target_tab.col1
(3) Digits
Either (2) or (3), but not both, may be omitted. If both are present, (3) may come before (2).