I have a question about the best practice data modelling specifically for Qlik Sense.In Qlik Sense the idea is to make the data easy for the end user to use in their own sheets. As In many systems, I need to build a model to suit lots of kind of related transactions. I have AR INvoices, Sales Orders, Credit Notes, AP Invoices, Purchase Orders, Inventory Adjustments, Payment Lines, etc.
The many different transactions all have common fields such as quantity and amount. In Qlikview I would most likely concat them all with different fact types but keep the same field names (quantity and amount). Then use variables with set analysis to create different types of info. Eg: vPOCost = 'Sum({$ <FactType={"PO"}>} [Cost Amount]). The in visualisations, just use the variables.
However, in Qlik Sense, we want to make it easy for the end user to build their own charts. So I was thinking of setting it up so the fact table has lots of similar measures: [PO Cost], [SO Cost], [AR Cost], etc. Then the user doesn't need to differentiate using set analysis.
Of course another idea is to use Master Items to build all these measures with inbuilt set analysis that the end user doesn't need to worry about. So the data model would simply have amount and quantity, but the template app would have lots of master items using set analysis to differentiate.
I don't think there is a right or wrong answer, but I am interested in the community's opinion.
