Report only MAX value in a query

Question

I have two tables I'm joining, a one to many relationship exists.
1 -> Many
A -> B
I want a few columns from table A in the report, along with a couple columns from table B.
Contents of table B are:
Product
Start_Date
Last_Activity_Date
Sample of data from table B:
PRIMARY,08/24/2007,08/24/2007
PRIMARY,08/24/2008,08/20/2009
PRIMARY,08/30/2008,08/26/2009
PRIMARY,08/01/2009,08/30/2009
How do I code the query so that only the last row is returned (the one with the highest Last_Activity_Date)? I originally thought I could use MAX(Last_Activity_Date) but that isn't working.
Discoverer's original code:
SELECT /*+ FIRST_ROWS */
     A.ASSOCIATE_NO Assoc_No,
     A.CLIENT_NO Clt_No,
     A.CLIENT_NAME Clt_Name,
     A.ACTIVE_EMP_CNT EE_Cnt,
     B.START_DATE Start_Date,
     ' ' AS Filler_Col
FROM A, B
WHERE ( (A.ID = B._ID)
AND (A.ID = B.ID(+)))
AND (A.PREMIER = 'Y')
AND ( (A.STATUS IS NULL)
GROUP BY A.ASSOCIATE_NO,
     A.CLIENT_NO,
     A.CLIENT_NAME,
     A.ACTIVE_EMP_CNT,
     B.PRODUCT,
     B.START_DATE,
     B.LAST_ACTIVITY_DATE
HAVING (B.PRODUCT_NAME(+) = 'PREMIER'
AND B.LAST_ACTIVITY_DATE(+) = MAX (B.LAST_ACTIVITY_DATE(+)))
ORDER BY 1 ASC, 2 ASC;
Discoverer's code isn't working the way I want it to. It returns all rows from table B.
How can I code it so that only the row with the highest Last_Activity_Date is returned?
Thanks for all your help,
Gary F.

Answer 1

Hi,
There are 2 ways of doing this:
You can create a rank calculation:
rank() over (partition by client_name, client_no /* and any other columns from A */ order by last_activity_date desc)
then create a condition where your rank calculation =1.
Or you can create max calculations for each column in table B and then don't have any conditions. So for your start_date calculation you would have:
max(start_date) keep (dense_rank last order by last_activity_date)
The second method is generally the best.
Rod West