Poor Man's SQL Pivot. List Questions as Columns and Answers per User in one row

Question

Current query:

SELECT order_id AS OrderNumber, ordName, ordLastName, question, answer 
FROM cart_survey 
JOIN orders 
    ON cart_surv         


        

Answer 1

This is known as a PIVOT there are two ways to perform this operation with a Static version or dynamic version.

Static Version is when you hard-code the values to become columns:

SELECT *
FROM
(
  SELECT order_id AS OrderNumber, ordName, ordLastName, question, answer 
  FROM cart_survey 
  JOIN orders 
      ON cart_survey.order_id=orders.ordID 
  JOIN survey_answers 
      ON survey_answers.id=cart_survey.answer_id 
  JOIN survey_questions 
      ON survey_questions.id=cart_survey.question_id
) x
pivot
(
  min(answer)
  for question in ([Type of Surgery:], [Month of Surgery:], 
                [Year of surgery:], [Current Ostomy System Brand:], 
                [Degree of Satisfaction:])
) p

A Dynamic Pivot, gets the list of the columns at run-time:

DECLARE @cols AS NVARCHAR(MAX),
    @query  AS NVARCHAR(MAX)

select @cols = STUFF((SELECT distinct ', ' + QUOTENAME(question)
                    from survey_questions
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')


set @query 
      = 'SELECT OrderNumber, ordname, orderLastName,' + @cols + ' from 
         (
            SELECT order_id AS OrderNumber, ordName, ordLastName, question, answer 
            FROM cart_survey 
            JOIN orders 
                ON cart_survey.order_id=orders.ordID 
            JOIN survey_answers 
                ON survey_answers.id=cart_survey.answer_id 
            JOIN survey_questions 
                ON survey_questions.id=cart_survey.question_id
         ) x
         pivot 
         (
            min(answer)
            for question in (' + @cols + ')
         ) p '

execute(@query)

Answer 2

The posted answers work but are clumsy and slow. You can do what I call parallel aggregation:

SELECT
     ID,
     SUM(case when question_id = 1 then 1 else 0 end) as sum1,
     SUM(case when question_id = 2 then 1 else 0 end) as sum2,
     SUM(case when question_id = 3 then 1 else 0 end) as sum3
GROUP BY ID

This will do one pass over the table instead of three and is very short. It is not a complete walk-through but you can surely adapt the concept to your needs.

Answer 3

This is the MySQL Version

SELECT o . * , 
q1.answer AS  'Type of Surgery:',
q2.answer AS  'Month of Surgery:',
q3.answer AS  'Year of Surgery:',
q4.answer AS  'Current Brand:',
q5.answer AS  'Degree of Satisfaction:'
FROM (
    SELECT DISTINCT ordID, ordName, ordLastName
    FROM orders
)o
LEFT JOIN (
    SELECT cs.order_id, a.answer
    FROM cart_survey cs
    LEFT JOIN survey_answers a ON cs.answer_id = a.id
    WHERE cs.question_id =18
)q1 ON o.ordID = q1.order_id
LEFT JOIN (
    SELECT cs.order_id, a.answer
    FROM cart_survey cs
    LEFT JOIN survey_answers a ON cs.answer_id = a.id
    WHERE cs.question_id =19
)q2 ON o.ordID = q2.order_id
LEFT JOIN (
    SELECT cs.order_id, a.answer
    FROM cart_survey cs
    LEFT JOIN survey_answers a ON cs.answer_id = a.id
    WHERE cs.question_id =20
)q3 ON o.ordID = q3.order_id
LEFT JOIN (
    SELECT cs.order_id, a.answer
    FROM cart_survey cs
    LEFT JOIN survey_answers a ON cs.answer_id = a.id
    WHERE cs.question_id =21
)q4 ON o.ordID = q4.order_id
LEFT JOIN (
    SELECT cs.order_id, a.answer
    FROM cart_survey cs
    LEFT JOIN survey_answers a ON cs.answer_id = a.id
    WHERE cs.question_id =22
)q5 ON o.ordID = q5.order_id

Answer 4

This is the MSSQL Version

select o.*, q1.[Type of Surgery:], q2.[Month of Surgery:], q3.[Year of surgery:]
    , q4.[Current Ostomy System Brand:]
    , q5.[Degree of Satisfaction with the fit and comfort of your Current Ostomy System:]
from (
    select distinct ordID, ordName + ' ' + ordLastName as [name] from dbo.Orders
) o
left join (
    select *, a.[Answer] as [Type of Surgery:] from cart_survey cs
    left join dbo.survey_answers a on cs.answer_id = a.id
    where cs.question_id = 1
) q1 on o.ordID = q1.[order_id]
left join (
    select *, a.[Answer] as [Month of Surgery:] from cart_survey cs
    left join dbo.survey_answers a on cs.answer_id = a.id
    where cs.question_id = 2
) q2 on o.ordID = q2.[order_id]
left join (
    select *, a.[Answer] as [Year of surgery:] from cart_survey cs
    left join dbo.survey_answers a on cs.answer_id = a.id
    where cs.question_id = 3
) q3 on o.ordID = q3.[order_id]
left join (
    select *, a.[Answer] as [Current Brand:] from cart_survey cs
    left join dbo.survey_answers a on cs.answer_id = a.id
    where cs.question_id = 4
) q4 on o.ordID = q4.[order_id]
left join (
    select *, a.[Answer] as [Degree of Satisfaction:] from cart_survey cs
    left join dbo.survey_answers a on cs.answer_id = a.id
    where cs.question_id = 5
) q5 on o.ordID = q5.[order_id]

Answer 5

This is called a pivot, where information in rows is used to determine the list of columns. This sort of query requires dynamically-computed SQL if done entirely in a query, and is usually better suited to client-side formatting instead (many tools call it a pivot or cross-tab query, SSRS calls it a matrix query).