Filtering maximum rank using rank function ?

Question

Hi ,
I have two table emp_changes(empno(number),operation_time(timestamp),operation(varchar))
and another table emp_metadata(tablename(varchar),max_time(timestamp))(it has got one row )
using this query :
    select empno,operation,RANK() OVER (PARTITION BY empno ORDER BY operation_time) "time" from emp_changes
where empno in (select empno from emp_changes where operation_time BETWEEN (select max_time from emp_metadata) AND sysdate)I am getting this output:
EMPNO OPERATION time
8009 INSERT 1
8009 UPDATE 2
8009 DELETE 3
8010 INSERT 1
8013 INSERT 1
9001 UPDATE 1
9001 DELETE 2
can anyone please tell me how to achive this like ,I want to view only distinct empno,operation with max rank in time like:
EMPNO OPERATION
8009 DELETE
9001 DELETE
Thanks ,
Deekay

Answer 1

Hi, Deekay,
Analytic functions (inlcude RANK) are computed after the WHERE cluase is appled.
To use the results of an analytic function in a WHERE clause, first compute the analytic function in a sub-query, then use the value in the WHERE clause of a super-query, like this:
select     empno
,     operation
from      (     -- Begin sub-query to compute time_rank
            select  empno
            ,       operation
            ,       RANK () OVER ( PARTITION BY  empno
                                           ORDER BY      operation_time     DESC
                              )           AS time_rank
            from    emp_changes
            where   operation_time     BETWEEN (
                                                select  max_time
                                  from    emp_metadata
                         AND      sysdate
     )       -- End sub-query to compute time_rank
WHERE     time_rank     = 1
;If you'd like to post CREATE TABLE and INSERT statements for your data, then I could test this.
While it is possible to see which is the highest number returned by RANK, it is easier to see which is the lowest number; the lowest number is always 1. That's why I used DESCending order.