Duplicate rows based upon condition

Question

Oracle 11.2.0.1
Windows
create table regsales (billno number,itemno number,paymode varchar2(10));
insert into regsales values (12345,10,'cash');
insert into regsales values (12345,11,'cash');
insert into regsales values (12346,11,'cash');
insert into regsales values (12347,10,'credit');
insert into regsales values (12348,14,'cash');
insert into regsales values (12348,15,'cash');
insert into regsales values (12345,12,'cash');
insert into regsales values (12349,10,'cash');
insert into regsales values (12345,10,'credit');
insert into regsales values (12350,11,'cash');
insert into regsales values (12351,12,'cash');
insert into regsales values (12352,11,'cash');
insert into regsales values (12350,11,'credit');
Required Output please :
    BILLNO     ITEMNO PAYMODE
     12345         11 cash
     12345         11 credit
     12350         10 cash
     12350         10 credit
i.e. in which same billnos same itemno has been sold in cash and credit.
Thank you.

Answer 1

Hi,
So, you need to know how many different paymodes there are in a group of rows.  That sounds like a job for COUNT (DISTINCT ...)
Here's one way:
WITH   got_cnt      AS
     SELECT  billno, itemno, paymode
     ,       COUNT (DISTINCT paymode)
                 OVER ( PARTITION BY  billno
                        ,             itemno
                      )  AS cnt
     FROM    regsales
SELECT    billno, itemno, paymode
FROM      got_cnt
WHERE     cnt  = 2
ORDER BY  billno, itemno, paymode
Can there be other paymodes besides 'cash' and 'credit'?  If so, the same basic idea will still work, but the details may be a little messier, depending on your requirements.