Could someone clarify my understanding?
In the daily ETL, normally fact table is loaded incrementally. For the dimension table, it is sometimes better to drop the whole table and reload. There are two issues -
1. When reloading the dimension table, the surrogate keys may be different for the same business key. Therefore, if any reports rely on surrogate keys, it may be wrong.
2. If the dimension is fully reloaded, when process the cube, I need to process Full on the dimension and all the related fact tables. I can't do Process Increase on those fact tables.
Is this understanding right? What is the normal way to think about these?
Thank you