I'm using GIT for QlikView development and deployment. We have a self-hosted Gitlab instance where each project is a repository. We work with a predefined template, that's setup in a specific way so you can start right away with a new dashboard which hooks into our data sets and gives you standardized features without having to start from scratch.
We switched to GIT after our ETL layer became too big to manage in development/deployment scenarios. Push/Pull is now our deployment method of choice. After making the change, I hardly ever touch the scripts inside QlikView anymore. Instead, the template is set up so that it will find find and load scripts in a specific subdirectory in it's own project folder. Data sources are found through variables loaded from config files.
We make use of config files extensively. It allows to build "engines" in QlikView script that do specific tasks with a very limited amount of code. As an example, the piece of code that manages the extraction of the SQL data across all our databases on different hosts is only 65 lines of code. We have a library with often-used functions t
Working this way has enabled us to manage our code base better, to have more freedom of working with the editor of choice (I use Visual Studio Code for the moment), to better handle deployments, and to cut down debugging time to the minimum. Using this method, we're managing an ETL script code base of 14.000 lines which are all loaded through a single qvd file that never has to be touched.
Another feature offered that I find extremely useful is forking and merging. It allows for different branches of the same codebase, allowing you to make customized solutions quickly, without rendering individual changes oblique and hard to merge back into your main project.
.png "Luminary Alumni"){kind=icon}
