Top N rows in a group by using row_number function | SQL

Let say I have this below data in my table for instance:

and my desired output is as below: Group it by WONUM but also I want to see the 2 records of each group.

Solution: (db2)

If I use group by clause based on wonum, then we will see only 2 records from wonum, 

if we group by WONUM, STATUS even then changedate will not display and we will not get our required output. 

Therefore, we need to group them and then count rows in each group , and based on the count we can apply any condition. 

let's do it using ROW_NUMBER() function:

ROW_NUMBER()

          OVER ( [partition_clause optional] order_by_clause)

 let's write our query:

SELECT *

from

(SELECT wonum, STATUS, changedate

    , row_number() 

    over (PARTITION BY wonum ORDER BY changedate) AS rank

  FROM WOstatus w

WHERE  W.WONUM IN ('42451047', '37778572') 

--GROUP BY wonum, status,changedate

) count

WHERE  rank <=2

Explanation:

we added a function (run_number() over) and make a partition/group by wonum ) give it a name as rank. 

2.  make this query as a child of another FROM clause ,

3. and in WHERE clause we put our condition to display only selected records because now we have a new column as rank. 
 

The same thing can be in Excel: 

Excel: Increment Numbers Based on Another Column(Criteria) when Value Changes | Windows