I'm relatively new to QlikView and have come into a new job where an existing implementation of a Qlikview solution is ingesting 7GB's of Monthly data from a SAS platform with 200+ attributes/columns for self-service reporting. The result has been poor Qlikview query performance and negative end users sentiment towards the QlikView product.
Assuming I cant do anything about reducing the number of measures and dimensions that have to be available for self-service reporting solution; how does one best utilise QlikSence to optimise performance? I can't help but feel the problem lies with the manner in which the solution was implemented more than the tool itself.
Can someone please send me some reference material on how a Qlikview self-service solution should be designed and how to navigate around common pitfalls associated with large data sets required for self-service solutions?
Additional Questions:
1. Can one break up the size of the dataset and send it as parts rather than one giant CSV?
2. Can one use ODBC instead of CSV? I'm being told that by the team that delivered the solution that a single CSV was the only way they could load the data and that loading the equivalent dataset via an ODBC was taking 7-8 hours...they've stated that they have used the latest ODBC drivers.
3. Can Qlikviews flat file be broken down to smaller data sets on the Qlikview back-end before being presented to the dashboard that requires only a subset of its data; or does every dashboard in a Self Service solution share a single superset?
4. What additional details would you need to help troubleshoot the cause such a problem? And at what point does one identify the cause of the problem being the infrastructure, not the way the solution was built?
