QlikTip #25: How do you organize your QlikView projects in your file system?

In the last three years I have quite often changed my standard concept of structuring my (customers’) QlikView projects in the file system. But for about 1 1/2 years “my” concept seems to be stable :-):

My favourite way of organizing the QlikView projects in the file system

Some hints:

  • In 0_Scripts all included scripts are saved
  • 1_Resources includes all Excel-files, settings files and other resources (eg. images, pictures, CSV-files, …) to be used within the load-script.
  • Under 2_QvdGenerators I place all my QlikView files generating QVD-files. All the QVD-files are placed within the subfolder “QVD” (and sometimes certainly subfolders of QVD)
  • In 3_DataModell I save my QlikView-file generating the datamodell, putting all QVD-files together to a nice datamodell.
  • In 4_Application the final result of my application is located, just loading the data modell with a BINARY LOAD and adding the user-interface to the available data.

Important notes and advantages of this approach

  • This structure does only work, if you have relative paths enabled. If doing so you can easily copy the development solution to the production folder or even to the production-system on another server.
  • Obviously this is just a standard structure, in smaller projects some of the folders would not be used, in larger projects I would add some sub-folders, but in general the structure remains the same …
  • When working in large projects (e.g. applications with 50 GB) you can easily work with a smaller datamodell (only a subset of the final dataset) and create your user interface without having to save each time a 50GB large file … for doing so just change the BINARY LOAD in your development environment and change it again to the large file in PRODUCTION ENVIRONMENT …
  • You can easily put this structure to your production environment just via copy&paste

How do you structure your QlikView projects?

I am interested in your personal way of organizing the projects.
Please use the comment functionality below to post your ideas.

Bookmark and Share

. Bookmark the permalink. Both comments and trackbacks are currently closed.

15 Comments

  1. Ralph G.
    Posted October 8, 2010 at 15:34 | Permalink

    We introduced this organization of our QlikView Projects when Stefan Walther came to us as a consultant back in November 08. Well, the only thing I can say that we’re still doing it and that it “works” just perfect for us!
    We only added another Step, 5_Reporting, which contains all the templates and designs for our QV-Excel reports. Also we have an “_Archive” folder for parking files before they get deleted.

    I highly recommend using this structure, or at least the basic idea! The most awesome fact is that setting the development environment to production is a matter of seconds/minutes (depending on the size of the folder to copy)

  2. Pablo
    Posted October 8, 2010 at 16:10 | Permalink

    I’m curious about the creation of DataModel. You create a QVW with all tables associated and no charts, then use this QVW in a binary load ?

    • Posted October 12, 2010 at 15:06 | Permalink

      Hi Pablo,

      yes you are right (and I’ve added some further clarification to the post).
      Certainly this approach takes a little bit more resources and loading-time, but on the other side it makes development easier (when working with really large QV-files) …

      Hope this helps …

      Best regards
      Stefan

  3. Marvin Tejada
    Posted October 21, 2010 at 05:01 | Permalink

    Hi!

    After developing more or more projects/applications already, I’ve already found the need to address my project files organization better. This looks just like what I’d need. Could you please send me an actual project folder with sample files in it just so I’d know what goes into what.

    TIA!

    • Posted October 27, 2010 at 15:52 | Permalink

      Hi Marvin,

      sure; I’ll send you (or publish) a sample project … please just give me a little time to prepare the sample-project …

      Best regards
      Stefan

      • gus
        Posted May 8, 2012 at 13:23 | Permalink

        Hi Stefan,
        I’m also intrested in a real example if you have one. It’s jus easyer to understand the approach.
        we’re trying to use it for larger projectd and I want to use something it can remains in time.
        Can you please upload or send it this example?
        Thanks in advance.
        Gus

  4. Posted October 28, 2010 at 06:56 | Permalink

    A great structure! I’m curious what you like to do when a QVD is shared among multiple applications?

    • Posted October 28, 2010 at 10:45 | Permalink

      In this case I have a seperate project-folder (only with 0_Scripts, 1_Resources and 2_QvdGenerators) where I generate the shared QVDs …

      This does not break possibility of xcopy-deployment which is one of my main-goals …

      Best regards
      Stefan

  5. satin
    Posted January 6, 2011 at 13:32 | Permalink

    HI
    any one plz help me out ……….

    i want to create a intelligence system with the help of qlikview ,but my database is in .csv file so i am not sure how to use that.

    • Posted January 28, 2011 at 10:45 | Permalink

      Hi Satin,

      it does not matter at all which type of source you are having for your BI-application: databases, flat-files (csv, txt), webservices, and so on …

      Best regards
      Stefan

  6. Posted April 26, 2011 at 16:10 | Permalink

    This looks very useful! It spawned some questions:

    * Is the Production folders for actually running production work with production data (as opposed to e.g. just generate a production release of the application)?

    * How do you use this on QV desktop vs. QV server installations?

    * For clients with strict requirements & controls on separating Dev from Test from Production, how would you typically adapt this?

    * Slightly off topic – any thoughts on how to best package up QV application releases in environments which require a formal release process?

    Thanks for your very useful blog!
    Kristian

  7. Marcel Hug
    Posted June 16, 2011 at 09:56 | Permalink

    I most like the idea to have one and only one generic data model (“3_DataModell”) which can be used by all the reports. Setting up the “correct” data model is fundamental, and it should always be easily extendable with new data structures.

  8. Lars Jackalin
    Posted November 17, 2011 at 16:36 | Permalink

    Hi Stefan.I have just recently started with QV an am looking for a good way of structuring QV projects and your approach certainly seems very good. I have a little problem understanding it though. I split up a very small program and ended up with having nothing in the 0_Scripts folder, everything ends up in the 3_DataModel folder. Is the Scripts folder just used för backup or am I missing something? Did you btw find the time to prepare the sample project. It might help to clear up things for me.

    Brgds / Lasse J

  9. Posted January 27, 2013 at 19:04 | Permalink

    I have a very similar approach to this but it’s different in a few respects like I also have a security folder and template folder. check out my article at http://webofwork.com/index.php/2-uncategorised/84-qlikview-project-folder-structure

  10. Sherwin
    Posted November 19, 2013 at 10:01 | Permalink

    Hi, i have a problem of calling of my qvd files up from my folder path.

    My qvw is here..

    ..\Application Data\ QlikTech\ Documents\ Development \ Ascella\

    while my qvd is from this path.

    ..\Application Data\ QlikTech\ Documents\ Databases \ Ascella\

    and my code is like this.

    LOAD [Name],
    [Age],
    [Address];
    [Location];
    FROM
    [QVD\Personal_Info.QVD]
    (qvd);

    Hope you could help me.

    Sherwin