What are some best practices when working with Binary Load Qlik Sense documents?
What are some best practices when working with Binary Load Qlik Sense documents? What should they have and what should you avoid when creating / working with a Qlik Sense app that will be binary loaded into another app?
Binary reload on server side is in my experience a bit tricky. The apps in server storage are used by the server processes for housekeeping and reload tasks, which means your binary reload can fail if the file is locked for other processing. I would recommend that you reuse data through a QVD layer rather than a binary reload.
Hi @ToniKautto, Could you please explain more on "The apps in server storage are used by the server processes for housekeeping and reload tasks..." I am trying to understand more what you mean by this.
App file is locked by reload process during reload, so you would need to time reloads to avoid theses resource conflicts. The server will read and write to the app files at times, for example to optimize file content or set app meta data. These are interactions that you will not be able to predicts, which means that there is a high likelihood that your binary load will fail due to the app is used by other processes.
For a successful binary reload, you would have to load from a QVF file, rather than the apps in the app storage (which have no file extension).
QVD layer is a better strategy for reusing data model and structure among apps.
In our original implementation from Qlik Services, the architecture we were told to use/setup is like this:
We have a series of QVFs which load data from our source LOB applications, data warehouses, files, etc and store QVDs.
We have a second series of QVFs which consume these QVDs, apply our business definitions/semantic model, and then save a new set of QVDs
We have a layer of Data Model QVFs which consume these QVDs
We have a layer of UI Applications which each binary load a single Data Model QVF (mentioned above)
All of our reloads are linked together, so the binary load of the DM (data model) happens when that application's reload is complete.
Lately we have start to have a number of issues around corrupt QVFs, corrupt QVDs, etc. Are you all indicating that this setup is not done correctly? We have hundreds of apps setup this way, and the architecture was delivered by Qlik services.
We do the same thing you describe in QlikView and now want to see if we can do it in Qlik Sense as well, like you. I want to be able to binary load my Data Model.qvf (which has Section Access in it) and cannot seem to get it working.