A typical decode in where clause

Question

Hi to all,
Could you anybody please give response to my query
select ast.account_num,sup.billing_status_text
from ast,sup
where
(sup.account_status=ast.account_status
or sup.account_status='99999')
Here sup.account_status,ast.account_status are notnull columns..
In the above query duplicates are coming as sup table has 2 rows , one row with matched value for ast.account_status and other row is with sup.account_status= '99999'.
I need to get only one row like
if matched row corresponding to ast.account_status is there in sup , then i need to get ,sup.billing_status_text corresponding to that matched ast.account_status
otherwise i shd get sup.billing_status_text corresponding to sup.account_status='99999' value in sup table
Can we use decode here to remove duplicates....
How to solve this

Answer 1

Hi,
Welcome to the forum!
Whenever you have a question, you should post your sample data in a form that people can use to re-create the problem and thest their soluitons.
For example:
CREATE TABLE     sup
(     ACCOUNT_STATUS          VARCHAR2 (10)
,     BILLING_STATUS          VARCHAR2 (10)
,     SUPPRESSION_FLAG     VARCHAR2 (10)
,     BILLING_STATUS_TEXT      VARCHAR2 (10)
INSERT INTO sup (ACCOUNT_STATUS,     BILLING_STATUS,     SUPPRESSION_FLAG,     BILLING_STATUS_TEXT)
     VALUES     ('99999',          'BB',          '99999',          'Unbill');
INSERT INTO sup (ACCOUNT_STATUS,     BILLING_STATUS,     SUPPRESSION_FLAG,     BILLING_STATUS_TEXT)
     VALUES     ('99999',          'CA',          '99999',          'Unbill');
INSERT INTO sup (ACCOUNT_STATUS,     BILLING_STATUS,     SUPPRESSION_FLAG,     BILLING_STATUS_TEXT)
     VALUES     ('OK',               'CA',          '99999',          'Bill');
INSERT INTO sup (ACCOUNT_STATUS,     BILLING_STATUS,     SUPPRESSION_FLAG,     BILLING_STATUS_TEXT)
     VALUES     ('PE',               'CA',          'N',               'Unbill');
CREATE TABLE     ast
(     account_num     NUMBER
,     ACCOUNT_STATUS     VARCHAR2 (10)
,     BILLING_STATUS     VARCHAR2 (10)
INSERT INTO ast (account_num, account_status, billing_status) VALUES (12356566, 'OK', 'CA');
INSERT INTO ast (account_num, account_status, billing_status) VALUES (96785674, 'AB', 'CA');
COMMIT;It also helps if you post the correct results clearly:
ACCOUNT_NUM BILLING_ST
   12356566 Bill
   96785674 UnbillSo the problem is, when you have more than one output row for the same account_num, you only want to display one of them.
You can do that by ranking the rows, using the analytic ROW_NUMBER function to assign numbers 1, 2, 3, ... to each row, with a different set of numbers for each account_num, and then pciking only the rows that were assigned number 1. (This is called a Top-N Query .)
What makes one row more desireable than another? You said, in the case of account_num=12356566, that it is the row where sup.account_status='OK'. Why? Is that because
(a) 'OK' is always better than 'PE' or '99999', or is it because
(b) 'OK' is also ast.account_status
? I'm guessing that it is (b), but if I'm wrong, you can change the analytic ORDER BY clause below:
WITH     got_rnum     AS
     SELECT      ast.account_num
     ,     sup.billing_status_text
     ,     ROW_NUMBER () OVER ( PARTITION BY  ast.account_num
                         ORDER BY        CASE
                                   WHEN  sup.account_status = ast.account_status
                                   THEN  1
                                   ELSE  2
                                 END
                       ) AS rnum
     FROM      ast
     ,     sup
     WHERE      (     sup.account_status     = ast.account_status
          OR      sup.account_status     = '99999'
     AND     ast.billing_status     = sup.billing_status
SELECT     account_num
,     billing_status_text
FROM     got_rnum
WHERE     rnum     = 1
; The analytic ORDER BY clause is similar to the query ORDER BY clause in that, if there are two or more rows with the same values for the ORDER BY columns, there's no telling in what order those rows will appear. You can add more expressions to the ORDER BY clause as tie-breakers, if necessary.