Question

Suppose now I have a table T which is like this:

A|B
1|1 
1|2 
2|1 
2|2 
2|3 
3|1 
3|2 
3|3 

With select A, count(*) from T group by A we should get:

A|count(*)
1|2
2|3
3|3

Now I need to get the last two rows in the previous query — which have the maximum value of count(*). How can I do this in a single query?

Thanks.

Answer 1

WITH recordList
AS
(
    SELECT  A, COUNT(*) totalCount,
            DENSE_RANK() OVER (ORDER BY COUNT(*) DESC) rn
    FROM    tableName
    GROUP BY A
)
SELECT  A, totalCount
FROM recordLIst
WHERE rn = 1

• SQLFiddle Demo