versioning and automated documentation

[Anonymous]

Hello,

I need to implement version control and possibly automated documentation for qlikview and sense in a small coworking environment.

I would like to:

- have a procedure that can allow multiple people to work on a project with no fear of saving over each other's work, (development -> staging -> production )

- avoiding scattered "archives" folders

- a tool for genereting documentation

So far I found only old postsin the community (2014 or before) while I think there are several new tools we can choose from. Sadly I have no idea what to choose between them so if anyone want to share his experience in this matter is welcome

thanks

[agigliotti]

what do you mean with "terrific" ?

is good or bad ?

[Anonymous]

Terrific = Awesome, sorry for my bad English.

Full modern angular UI like Qlik Sense, easy to use and we love the integrated Complexity Analyzer.

[bjendrick]

I'm currently in the process of implementing the QDF - Qlik Deployment Framework, and using mostly external linked QVS files for code, and storing all of that within a Git repository.  That way, we can control versioning, allow for multiple user development, and have a distributed backup of our versions.

Granted, this doesn't address documentation, so I've been attempting to figure out something that can work within a Git repository to record our activities.

[Anonymous]

The remedy for that is to set your variables in the script.

We use GIT to manage our -prj folders in QV and it works very well. I can't quite understand how anyone can work without version control. 🙂

[steinrich]

Hi,

we have evaluated PlatformManager from in4bi

It could be ok but it is an invest to take and another dependency from a software vendor.

Some enhancements should be made to the software dialogs and it could be more training necessary for external developers.

We decided to use an organizational way with clear roles in conjunction with simple Rational Clear Case checkout / checkin and Jenkins for deployment.

Just evaluating git-lfs instead of git-annex because Clear Case can be non-future proof.

Nevertheless you have to train and organize documentation over and over again, put it as a task into your quality gateway.

Richard

[Anonymous]

Richard,

Are you putting the qvw binary in Clearcase or the xmls in the prj-folder?

[jp_golay]

Hi Richard,

Have you been able to source control "Qlik Sense" with Rational Clear Case checkout / checkin and Jenkins for deployment.

I would be interested to know how to?

Thanks, JP

[Anonymous]

Ooops, just realized this was a Sense thread. 🙂

[steinrich]

Erik,

after a long period of evaluation and conceptual work we decided for an organizational programmatic way instead of another tool or the "prj xml does not stick together what is was before approach":

We prepare our qvw files for version control with this simple 2 actions:

Action A:

1 click: remove all data: for qvw types like qvd_generator.qvw, datamodel.qvw, datamodel_with_secu.qvw

Action B:

Some clicks and remove data --> keep possible values: for qvw types which have frontend objects and layout for enduser focus: make some simple selections on level 0 of all dimensions to select a few data records, layout and "always one selected value-property" still exists afterwards. If you have alternate state you have to make the selection on default state.

It works quite well and reduces 1GB files to less 4MB for all our qvw files, the 4 MB binary is stored in Clear Case then.

Next idea is to work with git-annex or git-lfs to get rid of the actions.

What happen if we have to use Sense? not yet thought about it, do you have ideas? do we need a new concept

[steinrich]

Hi Jean-Philippe,

yes we do.

See my answer to Erik.