Question

I’m having some trouble finding out the proper way to approach this question.

Database tables:
Student(stdID[pk], stdName, stdSet)
Enrollment(stdID[pk], crsID[pk], year[pk], semester[pk], grade)
Offering(crsID[pk], year[pk], semester[pk], instrID)
Course(crsID[pk], crsTitle, creditHrs)
Instructor(instrID[pk], instrName, dept)

“Find all instructors who had taught, from 2006 to 2008, a given student A01234567. List the student ID, instructorID, and year. Do not duplicate rows.”

What I originally had was this:

SELECT DISTINCT stdID, instrID, year
FROM Student s
JOIN Instructor i ON Offering o (o.instrID = i.instrID)
JOIN Offering o ON Course c (c.crsID = o.crsID)
WHERE stdID = 'A01234567'
    AND date BETWEEN (2006 AND 2008);

However, this is not entirely correct. I have notes that it’s something to do with “missing 2 columns from Offering join”, but I don’t know what that means…

Answer 1

Conceptually, I would start with the Instructor table, since that’s what I want to return.

I’d join that to the Offering table, so I have all the Offering for each instructor.

(But we don’t really need anything from the Instructor table at all, since we have InstrId in the Offering table.)

Then I’d join that to Enrollment, to basically get all the students that were enrolled in each course offering.

Then I’d join that to Student. But again, we don’t really need anything from the Student table, we already have stdId in the Enrollment table.

The tricky part is that join between Offering and Enrollment. The join predicate will need to be on year and semester, as well as crsId.

So, to satisfy the stated requirements, only two tables need to be queried:

SELECT e.stdId
     , o.instrId
     , e.year
  FROM Offering o
  JOIN Enrollment e
    ON e.crsId = o.crsId
   AND e.year = o.year
   AND e.semester = o.semester
 WHERE e.stdID = 'A01234567'
   AND e.year >= 2006
   AND e.year <= 2008
 GROUP
    BY e.stdId
     , o.instrId
     , e.year
 ORDER
    BY e.stdId
     , o.instrId
     , e.year

---

There’s a peculiarity with the database design. Consider what happens when there are two or more instructors teaching a given course in the same year and semester.

But the student was enrolled in only one of those.

This query will pick up the instructors of both offerings.

It’s a problem with the query, but the given database design doesn’t give us any way to resolve the issue… because there’s no way to tell which instructor the student had.

The query satisfies the stated requirements, but the results will be a bit odd, showing a particular student had six different instructors for ‘Calc I’ in the fall of 2006.

Unless there’s something that’s eluding me here.

EDIT:

As Emily Litella would have intoned, “Never mind…”. There’s a unique constraint on the Offering table. There can be only one offering of a given crsId in a semester. So there’s no problem there. (Except that it’s peculiar that there would be only one offering of a given course in a given semester.)

The GROUP BY is not necessary. The constraints already guarantee that there will be no duplicate rows returned.

Actually, I think the GROUP BY is needed, because it’s possible that the student was enrolled in two separate courses led by the same instructor. The purpose of the GROUP BY is to eliminate any duplicate rows from the result set (as indicated in the specification.)