Hello all,
Well, this topic is one of the most discussed on the forum, I am still unable to have a clear idea of when using QVD load and when using resident.
I just won half of a loading time using qvd load instead of resident, both wihout any transformation on a November 2018 with 26 millions rows. This seems counterintuitive since qvd load means using the disk while resident uses the RAM and data already loaded. So I think of a few factors that may have an impact :
-Qlik Sense/View Version
-number of rows and columns
-speed of disk (ssd versus hdd?) and ram (DDRxxx)
-cpu?
-concurrent access to ram and disk?
-use of load * versu load field 1, field 2...?
-type of data (string, num...) and cardinality?
We all have experience when one is fasther than the other but we never had a clear explanation from Qlik to my knowledge 😞
As a reference :
https://qlikviewcookbook.com/2018/03/preceding-load-performance-update/
Bi Consultant (Dataviz & Dataprep) @ Business & Decision