1 Large Table or build table relationships

[mathew101]

Hi All,

I'm fairly new here.  My boss, (not a data man).  Wants our I.T. to build one massive data table with everything in it.

In my mind this is a setting things up for problems, What if a supplier name changes, what if there is an error in the data.

It's more effort to I.D. problems with a massive file then with updating smaller related tables.

Is there any resource you can suggest, to help highlight the pros & cons of both approaches? 

[dwforest]

Look up resources that define how to define a Snow-Flake data warehouse.  It's not required for Qlik, but makes it much easier.

You end up with a cube in Qlik, so it all depends on your data what works best and Qlik has amazing compression so repeated data is not stored.

[Anonymous]

I usually aim for a star schema for the qvf data model.

https://qlikcentral.com/2014/08/19/the-optimal-qlikview-data-structure/

Depending on number of data sources / table and data volumes then a QVD layer is good.  I usually use a QVD layer even for simple, tiddly things.  It also make things easier to build & to diagnose data issues, plus the QVD's can be reused in multiple dashboards as needed.

Three Tier Architecture

The underlying data engine / modelling is the same QlikView vs. Qlik Sense and most of the good resources available are old QlikView ones.

The Qlik Sense front end viz's are very different though.