Remove duplicates based on a condition

Question

Hi all,
I need help on a query to remove duplicates based on a condition.
E.g. My table is
FE CC DATE FLAG
FE1 CC1 10/10 FB
FE1 CC1 9/10 FB
FE1 CC1 11/10 AB
FE1 CC2 9/10 AB
FE1 CC2 10/10 FB
FE1 CC2 11/10 AB
I want to remove all duplicate rows on FE and CC based on the below condition :
DATE <MAX(DATE) WHERE FLAG='FB'
That means I want to remove the row FE1 CC1 9/10 FB
but not the rows
FE1 CC1 10/10 FB
and
FE1 CC1 11/10 AB
as only the row FE1 CC1 9/10 FB has date <MAX(DATE) WHERE FLAG='FB'.
Similarly I want to keep
FE1 CC2 10/10 FB
FE1 CC2 11/10 AB
but not
FE1 CC2 9/10 AB
Many thanks.

Answer 1

Hi,
Do you want to DELETE rows from the table, or just not show some rows in the output? Since you're talking about a "query", rather that a "DELETE statement", I'll assume you want to leave those rows in the table, but not show them in the output.
Here's one way:
WITH     got_r_num     AS
     SELECT     fe, cc, dt, flag
     ,     RANK () OVER ( PARTITION BY  fe_cc, flag
                      ORDER BY          dt       DESC
                    )          AS r_num
     FROM     table_x
SELECT     fe
,     cc
,     TO_CHAR (dt, 'fmMM/YY')          AS dt
,     flag
FROM     got_r_num
WHERE     flag     != 'FB'
OR     r_num     = 1
;if you'd care to post CREATE TABLE and INSERT statements for your sample data, then I could test it.
This assumes that the column you called DATE (which is not a good column name, so I called it dt) is a DATE, and that you are displaying it in MM/YY format.
This also assumes that dt and flag are never NULL.
If I guessed wrong about these things, then the query can be changed; it will just be a little messier.