to prevent a future problem with an application Qlik that is going to load a huge quantity of data, i'll like to know if it is possible to have an hybrid system.
During the presentation of the data some area, for example the product takes data from a qvd, other area, for example orders, takes data directly from the database.
In this way i can present all the data without load all from qvd, saving some terabyte and don't strain on the resurce of the server.
If it is possible do you have some example or documentation?
The second question is if it is possible to have a qvw that present generic data and, for example when i filter some data in a table or chart, automatically switch to another qvw that usa the filter set before and show detail data.
you're right Direct Discovery is still very limited. We hope to see more functionality in the future. Another downturn of DD is the initial load of all dimension occurances from the given fact table. It could take at least hours in Big Data enviroments to gather this data. To solve this we could think about query rewrite in the background but this is another story.
One solutions would be to segment the data in several QVWs and to address or jump the right segment in the right QVW document from a central generic QVW (document chaining). You can open the related document by link or button depending from the current selection.
Another solution is a 2-step approach where you would have a similar generic QVW and after a selection of the data you want to analyse to start a reload of a second QVW with this filter. Btw. we use this kind of scenario in Big Data environments where you never would be able to load all data into QlikView.