Dynamic Pivot Columns with related tables in Sql Server - sql

Question

I am attempting to use a Pivot to dynamically create columns and am not getting the proper results I am looking for. I need to show all possible answers in its own column for each question (row), and there are multiple questions in a course. How can I generate each possible answer as a column name, and filter by courseId and the questionId for each row? The number of possible answers varies based on the question. Should this be done via cursor instead of pivot?
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX),
#courseID float(24) = 1
SET #cols = STUFF((SELECT distinct ',' + QUOTENAME(questionID)
FROM answers
INNER JOIN questions ON questions.questionID = answers.questionID
WHERE questions.courseId = #courseID
order by 1
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
, 1, 1, '');
set #query = 'SELECT * from
(
select Q.courseId, C.courseName, Q.question, Q.questionID, A.answer
from questions Q
inner join courses C ON Q.courseId = C.courseId
inner join answers A ON A.questionID = Q.questionID
where (Q.courseId = ''' + Str(#courseID) + ''')
) x
pivot
(
max(answer)
for questionId in (' + #cols + ')
) AS p'
execute(#query)
When running the above I am getting:
[1],[10],[11],[12],[13],[14],[15],[16],[2],[3],[4],[5],[6],[7],[8],[9]
As the generated columns which are the questionIds and not 'Answer 1, Answer 2 ... etc', and they are not displaying in order.
For the sake of brevity here are the related columns in the database tables:
Courses
courseId
Questions
questionId question courseId
Answers
AnswerId questionId answer
Any help is appreciated.

Answer 1

You didn't provide a many details on your table structure, but if you want the answers as the columns i.e. Answer1, Answer2, etc then you need to create the columns using row_number() against the number of answers per question similar to the following:
SET #cols = STUFF((SELECT ',' + QUOTENAME('Answer'+cast(seq as varchar(10)))
FROM
(
select row_number() over(partition by q.questionid
order by a.answer) seq
from answers a
INNER JOIN questions q
ON q.questionID = a.questionID
WHERE q.courseId = #courseID
) d
group by seq
order by seq
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
, 1, 1, '');
This creates a sequence for each answer per question and then that sequence number is used to create the new columns names. Then you will alter your PIVOT code to:
set #query = 'SELECT *
from
(
select Q.courseId, C.className, Q.question, Q.questionID, A.answer,
''Answer''+cast(row_number() over(partition by q.questionid
order by a.answer) as varchar(10)) seq
from questions Q
inner join classes C
ON Q.courseId = C.courseId
inner join answers A
ON A.questionID = Q.questionID
where (Q.courseId = ''' + Str(#courseID) + ''')
) x
pivot
(
max(answer)
for seq in (' + #cols + ')
) AS p'
execute(#query)