Question

code updated:
Added [score] column to #left table. The join must pull student name and score from #left table. Any one score is fine but it should not pull duplicate student_id. I am thinking cursor can do it?

---

I am running this report. The problem is similar to the the script that I created. To sum up, I have two tables lets call it #left and #right. The #right table is the key one which contains the amount. But the #left table contains the agent name which I must retrieve too.

create table #left (
id int not null primary key identity,
student_id int, name varchar(20),score int)

insert into #left values( 1, 'James',10)
insert into #left values( 2, 'Parker',20)
insert into #left values( 3, 'Smith',30)
insert into #left values( 4, 'Rog',40)
insert into #left values( 1, 'James',50)
insert into #left values( 2, 'Parker',60)

create table #right (
id int not null primary key identity,
student_id int,
amount decimal(5,2)
)

insert into #right values (1,5.25)
insert into #right values (3,7.25)
insert into #right values (4,3.25)
insert into #right values (1,5.25)

The problem here is I want to join the two tables on student_id but as you can see #left table contains James and Parker 2 times with the same id. Lets assume this was database mistake.

A similar problem can exists in the #right table. But that is not a mistake. It simply means the #left student was associated with two payments (even if the two payments are the same).

I want to remove dups from #left table and keep the dups in the #right table.

I came up with this query, which works

/* query 1 */
select student_id, amount from #right R
where exists
(
select student_id from #left L
where R.student_id = L.student_id
);

But the problem here is I also need to pull the student name from the other table. Is there a way to do it. It can be any syntax but efficient query is desirable.

Answer 1

This should work:

Solution to the original problem

Where #left table looked like this

create table #left (
id int not null primary key identity,
student_id int, name varchar(20))

insert into #left values( 1, 'James')
insert into #left values( 2, 'Parker')
insert into #left values( 3, 'Smith')
insert into #left values( 4, 'Rog')
insert into #left values( 1, 'James')
insert into #left values( 2, 'Parker')

Query: 1

SELECT A.student_id, A.amount, B.name
FROM #right A
LEFT JOIN (SELECT DISTINCT student_id, name FROM #left) B
ON A.student_id = B.student_id

UPDATED to the modified criteria Well, if any score is fine, then you can do this:

Query: 2

SELECT A.student_id, A.amount, B.name, B.score
FROM #right A
LEFT JOIN ( SELECT student_id, name, MIN(score) score
            FROM #left
            group by student_id, name) B
ON A.student_id = B.student_id