Solving circular connectivity between tables loaded from DB

[Googoogaga]

I’ve been trying to rework a data model for an Qlik Sense app that’s built on data of a creative site. The data comes from Postgres and Matomo.

The problem I keep coming to is how to connect the tables from Postgres correctly inside Qlik. I have a Users table, a Canvases table, an Objects table and a Shares table. It gets tricky because of the several relation points between the tables:

user_id -> canvas_creator, canvas_last_editor (inside Canvases), object_creator, object_last_editor (inside Objects), sharer_id, recipient _id (inside Shares)

canvas_id -> object_canvas (inside Objects), shared_canvas (inside Shares)

You might notice that not only does this create circular reference, it also means the same users_id field is meant to connect twice to the same table, in several cases.

How would you solve this? I’ve been reading a lot of posts in here and also tried coming up with several solutions and models to overcome this, but never quite got the result I wished for.

[Vegar]

I suggest adding an user table per user dimension  in your data model. (keep the naming unique)

Then join the three Canvases, Objects and Shares using the canvas_id. 

See mockup screenshot from QV below. 

[Googoogaga]

Hi! That’s an interesting solution, the problem I have with it is that I lose the connection between user columns. So I don’t have a unified user field that I can aggregate and filter with. Do you know how I would solve that?