In general it's not possible because Qlik isn't row-level oriented like the classical sql-databases else it's column-oriented. This means data couldn't be simply append like you would do it manually as an example by copying a log-record from one log-file and appending at the bottom of another log-file and saving this task. Qlik organized the data within system-tables (one for each field which stored only distinct values) and data-tables which just contain a binary pointer to the system-table.
For each adding/removing from data these system-tables and the pointer must be re-created. There is an exception by using partial loads but it has some restrictions in regard to the uniqueness of the values. Depending on your data and requirements this may be a problem or not. Beside this I doubt that you would significantly save load-times with such an approach. Nevertheless you may try it:
Scripting/ScriptPrefixes/Add.htm
Therefore I think my lastly made suggestion of reloading the dashboard binary and adding then the new data with an optimized qvd-load will be the fastest way to refresh your dashboard. By a well-built data-architecure + datamodel and an appropriate Server it should work quite well for your mentioned parameter of 200 GB of data (are this really 200 GB in Qlik and by optimizing the data-structures ?) and a refresh time-frame of 15 minutes.
- Marcus
.png "Former Employee"){kind=icon}
