SQL data query request - mysql

Question

I'd appreciate your help with an SQL problem.
I have some student quiz score data in an SQL table and I wish to write a query to extract the information that I want. Candidates can attempt the tests as many times as they wish. Ideally, for each candidate, I wish to find out their highest percentage score on each of the tests. And I wish to get an average percentage of their highest percentage score on each test. Many of the candidates will not have done all of the tests. For example, candidate 1's highest scores on tests 1, 2 and 3 are 50%, 100% and 0%, leaving an overall average of 50%.
The table is named resultsets. The relevant column titles names are: Candidate (this is the student ID number), QuizName (the title of each quiz), and PercentageScore. It looks like this:
Candidate | QuizName | PercentageScore
---------------------------------------
1 | Test1 | 25
1 | Test1 | 50
1 | Test2 | 100
1 | Test3 | 0
2 | Test1 | 50
2 | Test1 | 100
3 | Test3 | 75
I'm hoping to get a table that looks something like this:
Candidate | Test1 | Test 2 | Test 3 | AveragePercentageScore
---------------------------------------
1 | 50 | 100 | 0 | 50
2 | 50 | 100 | 0 | 50
3 | 0 | 0 | 75 | 25
(Thanks Jain) I'd like to know the SQL command that I should enter.
Thank you!

Answer 1

Aside from being a beginner, it would be good to get a handle on basic table / database structures, relationships, use of primary / foreign keys and especially data normalization.
As for learning queries, I have seen other people utilize SQL Zoo as it has sample data and covers samples of how to look for certain things that require different query, joins, left-joins, aggregates etc.
All that said, sometimes it makes things easier if you can understand queries based on YOUR data, not some generic sample database that you have no context on its application to your data.
With all that said, I will help you get started. You need aggregates (min, max, avg, count) that are typically applicable based on a "GROUP BY" column(s). In this first case, you want to find "for each candidate" (the group by), and each "QUIZ" for that CANDIDATE (also part of group by), you want the highest test.
SELECT
Q.candidate,
Q.quizname,
MAX( Q.PercentageScore ) as HighestScore
from
YourQuizTable Q
group by
Q.candidate,
Q.quizname
Will result in the following.
Candidate QuizName HighestScore
1 Test1 50
1 Test2 100
1 Test3 0 (a legit score on file)
2 Test1 100
3 Test3 75
From that, you could create a pivot. Now, different sql engines have different pivot syntax, but to better see on these specific quizes posted, I will be doing a hard-coded pivot. Since the pivot is derived (uses the first query as the basis), the first query IS the basis of the pivot.
SELECT
smry.Candidate,
if( smry.quizname = 'Test1', smry.HighestScore, 0 ) as HiTest1,
if( smry.quizname = 'Test2', smry.HighestScore, 0 ) as HiTest2,
if( smry.quizname = 'Test3', smry.HighestScore, 0 ) as HiTest3,
AVG( smry.HighestScore ) as AvgTest
from
( SELECT
Q.candidate,
Q.quizname,
MAX( Q.PercentageScore ) as HighestScore
from
YourQuizTable Q
group by
Q.candidate,
Q.quizname ) smry
group by
smry.Candidate
The "IF()" is applied as each row is attempted, and each row will only ever have an instance of 1 quizname, it would only be either "Test1", "Test2" or "Test3". IF it IS that proper test, then grab the highest score as the basis to show in that column result. The last column is a simple average.
The final group by is now keeping them per candidate, but this time WITHOUT the group of the quiz as the inner query had.