An effective method is a CROSS APPLY VALUES UNPIVOT (described in my signature links):

WITH Students (id, studname) AS

(

SELECT 1, 'x'

UNION ALL SELECT 2, 'y'

UNION ALL SELECT 3, 'z'

UNION ALL SELECT 4, 'a'

UNION ALL SELECT 5, 'b'

),

Markslist (id, maths, physics, English) AS

(

SELECT 1, 50, 60, 70

UNION ALL SELECT 2, 70, 60, 40

UNION ALL SELECT 3, 50, 80, 70

UNION ALL SELECT 4, 50, 100, 70

UNION ALL SELECT 5, 90, 60, 70

)

SELECT studname, [subject], score

FROM

(

SELECT studname, [subject], score

,rn=ROW_NUMBER() OVER (PARTITION BY [subject] ORDER BY score DESC)

FROM Students a

JOIN MarksList b ON a.id = b.id

CROSS APPLY

(

VALUES (maths, 'maths'), (physics, 'physics'), (English, 'English')

) c (score, [subject])

) a

WHERE rn=1;

Using ROW_NUMBER() instead of GROUP BY will even tell you the "stud" whose name got the high score.

If there's a tie for high score and you want to show that, use RANK() instead of ROW_NUMBER().