Question

I have survey statistics – for each answer (there is an arbitary number of answers) – I need to know the age group and the income group. (E.g. – answer “a” was answered by 23 people of ages 30 to 35, and by 11 people with income level of 10,000 to 15,000. and by 7 people of ages 30 to 35 with income level of 10,000 to 15,000 (this can be inferred by the first two requirements. Those groups are predefined).
This is actually a “cube” With:
1st dimension – answer id, 2nd dimension – age group, 3rd dimension – income group.
How should I represent it in the DB (using tables)? Can you make a generalization for cases where I need more than 3 dimensions of data?

Thank you all in advance.

Answer 1

The example has five dimensions: User, Question, Answer, Survey, and Demographic. The Demographic table is here to freeze the income level and age group at the time of the survey — the user can change income and age over time. The CurrentDemographicID in the User table is pre-calculated before the survey, and is later copied to the DemographicKey of the UserQA table.

select 
      AgeGroup
    , IncomeLevel
    , count(1)   as NumberOfPeople
from UserQA      as f
join Question    as q on q.QuestionID    = f.QuestionID
join Answer      as a on a.AnswerID      = f.AnswerID
join Demographic as d on d.DemographicID = f.DemographicID
join Survey      as s on s.SurveyID      = f.SurveyID
where QuestionFullName = 'specific question here'
  and AnswerFullName   = 'specific answer here'
  and SurveyFullName   = 'specific survey here'
group by AgeGroup, IncomeLevel
order by AgeGroup, IncomeLevel ;

Also, take a look at this question/answer.