How make possible Joins with group by clause - sql-server

Question

I have two tables, table1 has 2 columns as
id name
1 Amal
2 Varun
3 Sari
table2 has 3 columns as
id Subject marks
1 Maths 80
1 Malayalam 75
1 History 45
2 Maths 90
2 Malayalam 85
2 History 50
3 Maths 88
3 Malayalam 75
3 History 80
My question is to find the names who has the maximum mark for each subject (Subject wisw topper) the resultant table have to includes the fields name subject and marks
I tested with the following query
SELECT
table1.Student_Name, (table2.subject), max(table2.Marks_obt)
FROM
table2
INNER JOIN
table1 ON table2.stud_id = table1.Student_ID
GROUP BY
[Student_Name], table2.Subject
HAVING
MAX(Marks_obt) IN (SELECT MAX(Marks_obt) AS total_marks
FROM table2
GROUP BY subject)
In SQL Server 2008, but I got the result as
name subject
Sari History 80
Varun Malayalam 85
Amal Maths 80
Varun Maths 90
how I get the topper of three subject with these manner?

Answer 1

You can use ROW_NUMBER() :
SELECT s.subject,s.name,s.marks
FROM(
SELECT t1.*,t2.subject,t2.marks,
ROW_NUMBER() OVER(PARTITION BY t2.subject ORDER BY t2.marks DESC) as rnk
FROM Table1
JOIN Table2
ON table2.stud_id = table1.Student_ID) s
WHERE s.rnk = 1

Answer 2

Use Rank You will not miss any people for example like two people will get same highest marks in same subject. And if you want query to find 2nd highest marks or 3rd highest use Dense_Rank() function even dense_rank() also works for finding 1st highest. For More cilck here
SELECT NAME,
SUBJECT,
MARKS
FROM (SELECT NAME,
SUBJECT,
MARKS,
rank()
OVER(
PARTITION BY [SUBJECT]
ORDER BY MARKS DESC) RNO
FROM #TABLE1 T
JOIN #TABLE2 T2
ON T.ID = T2.ID) A
WHERE RNO = 1

Answer 3

CREATE TABLE #table1
(Student_ID INT,
Student_Name VARCHAR(20))
INSERT INTO #table1
SELECT 1,'Amal'
UNION
SELECT 2,'Varun'
UNION
SELECT 3,'Sari'
CREATE TABLE #table2
(
stud_id INT,
[subject] VARCHAR(20),
Marks_obt INT
)
INSERT INTO #table2
SELECT 1,'Maths',80
UNION
SELECT 1,'Malayalam',75
UNION
SELECT 1,'History',45
UNION
SELECT 2,'Maths',90
UNION
SELECT 2,'Malayalam',85
UNION
SELECT 2,'History',80
UNION
SELECT 3,'Maths',88
UNION
SELECT 3,'Malayalam',75
UNION
SELECT 3,'History',80
/*Table 1*/
SELECT * FROM #table1
/*Table 2*/
SELECT * FROM #table2
/*Top Mark*/
SELECT [subject],
Student_Name,
Marks_obt
FROM(SELECT Student_Name,
[subject],
Marks_obt,
RANK()
OVER(
PARTITION BY [subject]
ORDER BY Marks_obt DESC) RowNum
FROM #table1 T1
JOIN #table2 T2
ON T1.Student_ID= T2.stud_id) AS data
WHERE data.RowNum = 1
DROP TABLE #table1,#table2

Answer 4

you can use a cross apply too like this
with maxi as (
select Subject, max(marks) maximark from table2
group by Subject
)
select * from maxi f1
cross apply
(
select top 1 f2.name from table1 f2 inner join table2 f3 on f2.id=f3.id
where f1.maximark=f3.marks and f1.subject=f3.subject
) f3
if multiple users are possible for a maxi mark, remove "top 1"

Answer 5

other solution with imbication:
with maxi as (
select Subject, max(marks) maximark from table2
group by Subject
)
select (select top 1 f2.name from table1 f2 inner join table2 f3 on f2.id=f3.id where f1.maximark=f3.marks and f1.Subject=f3.Subject) as Name, f1.*
from maxi f1