3

I'm new to QlikView and want to use some version control systems (such as Git) in order to track changes.

I am however not sure exactly how to do this, because at first sight it seems to me that the .qvm files contain both code and data. As a devloper this seems a bit odd, and if true a big limitation wrt source control.

Does anyone have a recipe for how to separate code from data in QlikView files?

Geir
  • 514
  • 4
  • 18

2 Answers2

6

In the folder where your qvw files sits create new folder named the same as the qvw but with -prj added.

For example if your qvw is named MyQVApp.qvw the folder should be named MyQVApp-prj. After this open your qvw and save it. After the doc is saved it will populate the -prj folder with bunch of xml and txt files. These files contains the qvw structure/layout (without the data) - the load script, sheets, sheets objects etc.

Then you can put the -prj folder under whatever version control you want. Every time the qvw is saved the files in -prj folder will be refreshed and you can commit your changes. And every time you open your qvw Qlikview will load the layout from the -prj files.

Also there is build-in integration with TFS and Subversion. For more info please read the pdf files from https://community.qlik.com/docs/DOC-3424

Stefan Stoichev
  • 4,615
  • 3
  • 31
  • 51
  • 1
    This works well with one exception: The values of variables are considered to be "data" and QlikView does not store the variables's values anywhere in the -prj. Workaround: Define the variables in the script. – Kevin P. Oct 22 '19 at 04:44
1

Stefan's answer is right, and I have done this with git for a while now. I have also made manual edits to the generated xml files and noticed that they are correctly treated by QlikView. Although I have not tested complicated merges from multiple developers, and merging XML is generally difficult to automate so I wouldn't be surprised if changes merged by git are poorly handled by QlikView.

Nameless One
  • 1,615
  • 2
  • 23
  • 39