thanks for swift reply holch!
I understand the confusion -- It's difficult to explain and maybe I'm getting it wrong, too. My aim is to avoid getting the result data stacked in columns but instead in rows - also it's good database practice to not duplicate when it's not necessary.
I made a survey using the setup you suggest, and it works, but it is not so practical, and can potentially create a data set that is very "wide" -- an example:
I wanted the result data organised like this:
Y1 Y2 Y3 Y4 Y5
X1 ** ** ** ** **
X2 ** ** ** ** **
X3 ** ** ** ** **
X2 ** ** ** ** **
.. instead of like this:
Y1 Y2 Y3 Y4 Y5 Y1 Y2 Y3 Y4 Y5 Y1 Y2 Y3 Y4 Y5 Y1 Y2 Y3 Y4 Y5
X1 ** ** ** ** ** -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
X2 -- -- -- -- -- ** ** ** ** ** -- -- -- -- -- -- -- -- -- --
X3 -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- ** ** ** ** **
X2 -- -- -- -- -- ** ** ** ** ** -- -- -- -- -- -- -- -- -- --
The solution would have been a simple array set if it weren't for my need to only have them fill out the sets of their selected Xs.
I'll make an attempt to exemplify it in a hands-on analogy:
You have 10 cars and you want the customers to express what they think about them with 100 questions for each car. The 100 questions are the same for all cars. Only thing is that they have all tried different cars, so you have them choose which ones they tried. Some tried one, others tried 7, or all of them.
I want to make a survey like this where the customers doesn't have to deal with the cars they haven't tried. This also without creating 10x100 of the same questions, leading to 1000 columns in the result data.
I want them to select the cars, and for each car go through the same questions, one car at the time. The result data would be 100 columns wide, with the car tag as the only difference (101 columns to be exact).
