1 Reply Latest reply: Dec 7, 2011 5:17 PM by John Witherspoon RSS

    server performance publisher

      Hi togehter,

      i have got a performance problem on our qlikviewserver and need advice.
      Meanwhile, our data model has grown up to 1.5 GB . The data model
      fed by three different systems. It consists mostly of the production data that compared over several years for qlikview-BI.

       

      As we carry out evaluations across all systems, we invite by the "binary" command the entire data model to feed any qlikview application.

      In our case we have approximately 20 applications that are each charged separately by the binary command with identical data model.

       

      You can imagine that it is quite to the hardware resources . Too long load times or extremely slow execution of the published applications .

       

       

       

      The idea is to divide he data model into small pieces and load only the data that are needed for the current application. Can I load in a QlikView application several small binaries ? Or how do I realize it so I do not always load the giant data model in every qlikView application ?

       

       

       

      Thank you very much !

       

      Regards,

       

       

       


       

       


        • server performance publisher
          John Witherspoon

          I would store each table as a separate QVD.  I would also join (or applymap) small helper tables onto your large fact tables so as to avoid having too many QVDs.  When loading from QVD, load only the fields you need, and where at all practical, use an optimized load.  QVDs are a very, very common approach.  So common that I worry I've misunderstood the question.