Distinct Key word in sql

Question

Hello Good folks,
I have a simple oracle view (that fetches data from mutiple tables).
There are approximately 7 columns in the view. I want the view to get unique data based on only 2 columns (not all 7 columns).
View looks like this now.
select distinct cusip, calendar-date,price,aa,bb,cc,dd
from view_aa
order by cusip.
view fetches Data like this.
cusip, calendar-date, price, aa, bb, cc, dd
01 Jan-01-2009 1000 10 12 8 5
01 Jan-01-2009 2000 10 11 8 5
02 Fed-22-2009 3000 11 12 6 4
But I want the view to get the following data only
cusip, calendar-date, price, aa, bb, cc, dd
01 Jan-01-2009 1000 10 12 8 5
02 Fed-22-2009 3000 11 12 6 4
Any advise/suggestions on how to get non-duplicate data?
Help much appreciated.
thanks
nath

Answer 1

Hi, Nath,
So you only want one row for every cuspid and calendar_date.
Is that one row
(a) some kind of combination (that is, an aggregate ) of all of the rows with that cuspid and calendar_date (if so, what?), or is it
(b) one of those rows, exactly as it is in the table? If so, which one?
I assume you want (b), where the row chosen is the one with the lowest price. If there are two (or more) rows with the same lowest price then, from among them, the one with the highest aa will be chosen. If there is still a tie, one of those rows will be chosen arbitrarily.
WITH     got_rnum     AS
     SELECT     cusip, calendar_date, price, aa, bb, cc, dd
     ,     ROW_NUMBER () OVER ( PARTITION BY  cusip
                         ,             calendar_date
                         ORDER BY        price
                         ,             aa          DESC
                       )     AS rnum
     FROM     view_aa
--     WHERE     ...     -- Any filtering goes here
SELECT     cusip, calendar_date, price, aa, bb, cc, dd
FROM     got_rnum
WHERE     rnum     = 1
;This is called a Top-N Query because you are choosing N rows (N=1 in this case) from the top of an ordered list.