Exam Post-Mortem

Hey all,

I must say it was difficult on my side - far more difficult than either of the past papers. I saw 3-4 people get up quite early from the exam, but judging from it, I doubt they finished it. Mostly everyone wrote from start to end.

My concerns are as follows:

• Lack of content covered
• Specific content had no "intro" marks, it was all complex
• Questions that fall into the "catch you out" category, I.e.: That INSERT statement would have failed without DEFAULT values

What do you all think?

It was quite tough. It should really have been a 3 hour exam

I felt all the hardest questions were at the front, and toward the back they got a bit easier. But stuff like definitions...dont ask me that.

Are you talking about the INSERT where it asked you to put R865 into the salary column (which is has a number datatype)?

I had a few thoughts regarding some of the questions:
1. The insert I mentioned above. Asked to put R865 into the salary column. Is that exactly what should be typed in, IE was string processing and To_NUM wanted? Or was doing the formatting yourself ok?

2. The bigger of the two tables which you had to provide the code to create, the one that was also in assignment 2. The model answer had nothing about the unique constraint on the job_card_num column, so were we supposed to replicate the model answer from the assignment or provide the REAL correct answer (IE. including the unique constraint).

3. My brain got suck in a repetitive loop for question 4 where I kept telling myself I cannot use avg(salary) in the where clause.

4. Some of the MCQ questions were tricky imho

5. Something else which escapes me at present. I shall comment later. EDIT: Oh ya, It never occured to me whether the salary as seen in the sample data was annual or monthly salary, and i dont beleive it was specified anywhere. Anycow I presume it was monthly, and multiplied by 12 in the question asking for annual salary

6. When osprey have spam prevention implemented?

1. It would be implicitly converted, but that was not the problem. The problem was the JOB_ID column, along with the limitation on the question of not specifying columns. They provided no job_id, and that column has a NOT NULL constraint. WHAM, insert error. I added "assumptions" saying that there was a DEFAULT value for it. The rest of the insert was easy peazy.

2. As per the exam, add the UNIQUE constraint. Doesn't mean a damn thing as to what the assignment said.

3. Well, you can't use a group function in a where clause. This was one of my "catch you out" questions. You needed to have a sub-query with 1 > (Subquery), and on top of it, return a grouped function. What I did was use the > ANY(Subquery) command. No idea what I'll get there.

4. Agreed, the wording was off, I.e.: "including the matching columns". I used INTERSECT instead of UNION or UNION ALL, as they made no mention of duplicates or non-matching columns.

5. Ok.

6. It's in the main announcements, check at the top.

1. I didnt notice that, but I choose to believe that was an error on the examiners behalf, as opposed to a sneaky "catch you out" situation.

2. I hope they dont use the assignment model answer

3. Editted above. I meant repetitive, not competitive. I knew I couldnt use avg in the where clause which kept me looping to step 1: where am i gonna put avg(salary)?. And FUUUUUU I thought of doing the 1 > method but decided against it in the end.
I put mine as (its wrong, but a noble attempt, I believe): select job_id, avg(salary) from employees group by job_id having avg(salary) = (select min(avg(salary)) from employees);

4. I used MINUS in all 3 set operator questions

5. I remembered, see above.

6. FUUUUU, forgot about the spam prevention.

Your #3 is not a bad attempt at all. It actually makes sense (which mine was complex).

The set questions I used the first 2 as MINUS and the last one INTERSECT, but those were also tough due to the column matching. I performed joins here to get around this issue.

For the salary, I also assumed monthly by the amount. If I earned R895.00 p/a I'd die. Did the same as you.

P.S: The spam prevention catches me almost every time.

The thing with sample data is that it can be a load of crap. Its not real. So I included a note saying I assumed the salary column was for monthly salary.

I wish they put less column joining stuff and more, i dont know, user management, subqueries, string processing (maybe).

WRT how I think I will do in the results, I think I will get a high mark.

I thought the exam was quite difficult... The MCQ were a little difficult.

I didn't get some of the queries... like the average salary question with the subqueries
It was a little confusing. I thought that maybe I didn't study hard enough, but I see that everyone else also found it quite difficult.

Lol, when I studied for this exam, i thought it would be a breeze- i guess I underestimated INF307. Wow,if I underestimated this one, I wonder how the other ones will be.. like the dreaded COS311 exam lol.

Im more concerned about COS351. I must take some vaseline into the exam room.

The exam wasn't what I had expected also but I tried my best, I guess I'll be kissing the distinction I was aiming for good bye. And COS311 I won't be writing it because I still don't get it so it will be a waste of time me writing it because am gonna fail it either way

--------------------------------------------------------------------
In a long journey, even a matchstick is heavy...

Wow, the exam did not go as expected. Completely different format to that of the past papers. Some high mark questions - 8 and 10 marks.
The first multiple choice question wanted 2 answers... that threw me a bit. My first multiple choice question with 2 answers in 5 years of writing UNISA exams.

I too found the exam got 'easier' as it went. Half an hour in as i was only on on question 2 (sec B ). I thought I was going to run out of time. I left question 2 and 4 (the high mark ones) to come back to later. At the end I had more than enough time.

Wasn't expecting any theory questions... "When does a trx start and end?" (4 marks). Lol, I think I got 1 mark for that after half a page of writing nonsense.

I also answered all the SQL Statement quesitons in pencil. They better be marked!

I thought I had done ok but the more I think about the less I am sure. I guess the problem is that when I run SQL queries for work I always take the best guess first and then tweak the results until the query is perfect which I guess makes one a little lazy...

Plus that insane question 2 of section 2 (the employees and their colleagues) threw me a bit right at the start.

Time for bed and then some mild panic for COS311!

Where did this paper come from? Why was this paper so much more difficult that last years papers.
Since when does a MCQ have 2 answers, that just threw me right from the beginning.

I doubt the MCQ question with two answers was intended, Section A was out of 10 marks, with 10 questions after all. There was probably something slightly wrong with one of the options.

The answers written in pencil have to be marked, there was no rule on the cover saying write in pen (I checked). Also, most programming courses (I guess this includes scripting) prefer exams to be written in pencil, because there is a lot of scratching out, editting etc done on code/scripts. I wrote most of my code in pencil too, until they ran out of lead (both of them ;< ).

I didnt mind question 2, there was a question almost identical in the practice questions. Or was it the assignment?

The biggest problem I had with the exam was the inconsistent wording and ambiguity in some of the questions. They refered to "employee id" as "employee number" at one point and then in another instance it is refered to as "employee id". Same with employee name - it was rarely made clear if they wanted first, last or both names and in seperate or the same column. They would also ask for different things in the same question. I remember them asking for employee name and salary (or something) at the start, and at the end of the question it states they want a specific order with a different column added or a specific sort order or something. I does not sound like a big problem, but I kept having to reread the questions to make sure that I got it right (or as close as what I though they wanted) and on a few occasions I just took a guess at what they wanted. It wasted a lot of time.

Hi all

Sho i regret having registered the module only this year, things change, to think i wasn't able to send the last assignment. what % do i need for supplementary i doubt it will be different. Getting this degree is not going to be easy.

Hi everyone

The exam paper was tricky. Not what i expected.

For the questions where they asked to use set operators, I didn't. My mind was all messed up. I only realized this when I was checking all my work and time was up by then.

What I'd like to know is will I get part marks for those questions (where you had to use set operators where I did not)?

@punk - I doubt it. The queries would have been so much simpler using JOIN's or otherwise, so it would probably be termed unfair to grant marks for doing it that way. You might get a mark, say, for the first select statement.

Hi

The exam was definitely harder than the samples we received for previous years, but I still don't think it was unreasonable or too hard. What did bother me was the ambiguity in some of the questions as posted by MotaBoy. A couple of questions did not clearly state which columns they wanted.

Also the question about the annual salary did not specify whether the given salary was annual or monthly. Also I remember a question in the practices in the text book which was the same as the one about the salaries, but the inserts was not in the actual employees table but a different table with the same structure as the structure provided in the question. In the exam the question stated insert the values into the 'employee' table, not the 'employees' which was the table given in the appendix. This once again left me with the question, should the insert statement be compatilble with the 'employees' table as described, or should the statement be comaptible with the given structure in the question as was the case with the same question in the practice in the text book.

Other than that I did not really have any problems with answering the questions, but then again I write much more complex queries on a daily basis as part of my job.

LouisD, I gotta ask: I too write much more complex queries than what was received in the exam. My point though, is I write it correctly - as in using the correct functions. When last did you use a SET operator to complete something that a simple INNER JOIN can achieve with half the work?

I see your point, and I agree with you. It frustrating to be told to do things in certain less optimal ways when doing assignments/exams, in this subject as in any other.

Unfortunately they need to test our knowledge of set operators and it's a lot easier (although not impossible) to ask the question using an example that does not need you to use set operators than to think of a legit example where using the set operators would have been the best solution.

To design questions where the best solution would always be the specific concept which they wanted to test would make setting the paper a lot harder (again, not impossible), and without specifying what concepts you should use for solving a query a lot of the questions can be answered by only using basic concepts and they would not be able to test our knowledge of more advenced concepts

We are in the fortunate(unfortunate?) position of doing this on a daily basis, therefore we can see that there exists a better solution, and luckily in a work situation we can use whichever solution we consider to be the best for the given scenario.

Fair response - was just pointing out my personal bug-bear.

Hi there

My take on this is that the exam was that it was very fair, but I have to agree some of the questions were very open ended and sometimes confusing as to what was actually needed. The other thing was that the mark allocation was sometimes a little extreme. The table create question was for like 10 marks? The format was also very different to last years paper.

The other thing was in the last question refered to the managers name, I could not find a manager table or a reff for manager first name on any of the tables? so I dont know if this question is correct, the other thing is judging from the mark allocation it could not have been a very complex question, with joins etc?

I thought there would be theory questions but not what we got. I think that it all depends on how this is going to be marked, if they are going to be very strict on syntax etc, we are going to have problems. but if they are looking at the basic logic etc, it may be OK,

So I am hoping to have done enough but I really think that this is all about how they mark this.

@robynsky
The managers name... the manager_id is an employee_id and therefore their details are on the employees table.

Yup .. i also found this exam to be quite a 'tuffy'.
I think the problem I have (and perhaps some of you'll will agree) is that i understand SQL can run complex queries to get the desired results BUT
when it comes to putting the code to paper the shit hits the fan... = 50% will do for me ..

Hey guys

I will be doing the following modules next semester: INF303D; INF305F; INF307H; INF308J; INF320E. Will u pliz email me any material you've got in any of the above modules.