SQL : SELECT all rows with maximum values and with WHERE condition also - sql

Question

I've table which look like this:
id data version rulecode
---------------------------
a1 1 100 x1
a2 1 100 x1
a1 1 200 x2
a4 2 500 x2
a7 2 200 x1
a6 2 500 x1
a7 2 500 x2
a8 2 150 x2
a9 3 120 x1
a10 3 130 x2
a10 3 120 x1
a12 3 130 x2
a13 3 130 x1
a14 3 110 x2
a15 3 110 x1
a16 4 220 x1
a17 4 230 x2
a18 4 240 x2
a19 4 240 x1
..........................
..........................
Now I want only those rows which has maximum version and data value as (1,2 and 4)
When I tried with dense_rank(), I am getting only those rows which have 1 value from data column:
SELECT * FROM
(SELECT *, dense_rank() OVER (ORDER BY version desc) as col
FROM public.table_name WHERE data in (1,2,4))x
WHERE x.col=1
Output:
id data version rulecode
---------------------------
a1 1 200 x2
My expected output:
id data version rulecode
a1 1 200 x2
a4 2 500 x2
a6 2 500 x1
a8 2 500 x2
a18 4 240 x2
a19 4 240 x1
Note: the value of data column is till millions.
Can someone help me out here to get the expected output?

Answer 1

You seem to want a PARTITION BY:
SELECT *
FROM (SELECT *,
DENSE_RANK() OVER (PARTITION BY data ORDER BY version desc) as seqnum
FROM public.table_name
WHERE data in (1, 2, 4)
) x
WHERE x.seqnum = 1

Answer 2

Using analytic functions:
WITH cte AS (
SELECT *, MAX(version) OVER (PARTITION BY data) max_version
FROM yourTable
)
SELECT id, data, version, rulecode
FROM cte
WHERE version = max_version AND data IN (1, 2, 4);
Note that we could have also filtered the data values inside the CTE. I will leave it as is, for a general solution to your problem.