I'm using QlikSense and we're ramping up for our first production release.
I have quite happily imported the data from BW and currently my model has three tables;
Table 1 - 2 Million rows
Table 2 - 1.5 Million rows
Table 3 - 0.5 Million rows
(plus some other, smaller tables).
Performance is becoming a bit of an issue, but I wonder if there is any benefit (either storage or speed) to the way the data is imported?
If Table 1 was a list of clients, each having an address, an account number, a gender, an age, etc. is it stored more efficiently inside Qlik if it is imported as a display attribute rather than a navigational attribute?
My thought is in a dense matrix like this (all objects have all attributes) it doesn't make a difference. Once imported into Qlik the internal representation is just columns of data - it doesn't care where they came from.
Is this correct?