TL;DR: how can I invalidate the cache of one cube?
I want to send a few notifications, based on one data set. The issue is that the number of notification depends on the data set, and I do not know in advance how many will be set. Dundas cannot handle this out of the box.
I worked around the problem by defining 1 cube and 1 notification. The cube reads its data from Postgres and sends 1 export. Outside Dundas, In Python, I wrote this kind of script using the rest API:
alldata = get_all_data() for d in alldata: update_postgres(d) send_notification()
In short, the cube and notification do not change, the underlying data does.
It works very nicely but for one issue: Dundas caches the data it receives from the cube. So my question is: how can I either invalidate the cache, either configure the cube (or metric set, or dashboard) to not use the cache?