Output not as per requirement

Question

Hi there,
I am running a query:
select distinct(state_code),count(*),hldy_str from hol,sol where cal_b2k_id=sol_id
and mmyyyy='122009'
group by state_code,hldy_str
order by state_code
is output is something like :
HAR 63 Y Y Y Y Y
HP 15 Y Y Y Y
JHK 44 Y Y Y Y YY
JHK 19 Y Y Y Y Y YY
JNK 4 YY Y Y Y YY
KAR 127 Y Y Y Y Y
KER 188 Y Y Y Y Y
KER 1 Y Y Y Y YY
MAH 1 Y Y Y Y
MAH 410 Y Y Y Y Y
MAH 1 Y Y YY Y Y
MAH 1
MEGH 4 Y YY Y Y YYYY Y
Now If you notice that there are two rows for JHK, 3 for MAH what I want is to have only one row in which high value for the 2nd col should come
Pls help

Answer 1

Hi,
I see: you want one row per state_code, and the highest count for that state.
Do you want hldy_str? If so, which one, or do you want some kind of aggregate?
What if there's a tie for the highest count in some state?
Depending on your requirements, you may have to change this:
WITH     group_by_2_columns     AS
     SELECT       state_code     -- no need for DISTINCT when you select all GROUP BY columns
     ,       COUNT (*)          AS cnt
     ,       hldy_str
     ,       RANK () OVER ( PARTITION BY  state_code
                           ORDER BY          COUNT (*)     DESC
                           ,          hldy_str
                         )     AS rnum
     FROM       hol
     ,       sol
     where       cal_b2k_id     = sol_id
     AND       mmyyyy     = '122009'
     GROUP BY  state_code
     ,       hldy_str
SELECT       state_code
,       cnt
,       hldy_str
FROM       group_by_2_columns
WHERE       rnum     = 1
ORDER BY  state_code
;The query above includes the entire row that had the highest count for each state_cd.
If there is a tie, it takes all of the tying rows. Change RANK to ROW_NUMBER if you want only one of them.
Notice that the sub-query (group_by_2_columns) is almost exactly what you posted.
The main query just picks the top row for each state_code from your original results.
I hope this answers your question.
If not, post CREATE TABLE and INSERT statements for your sample data, and the results you want from that data.