Data Model Resident Loads

[schneider2037]

I'm working through overnight performance hits on the QV Server.

One item that I'm seeing is that our Data Model QVWs often have dozens or more of Resident loads against underlying QVDs as well as large SQL outputs from our ERP.

QUESTION: do all of these Resident loads within each QVW really have that much of a performance impact?

Thanks! Jonathan

[MK_QSL]

You can try both

Resident Load

Preceding Load

to find out with one is more effective for performance point of view.

Read below blog..

The Cost of Preceding Load | Qlikview Cookbook

[ychaitanya · Accepted solution]

i would suggest to load the data from QVDs wherever possible than compared to using the Resident Loads

and implement exists() with combination of INLINE to implement filters

Thanks

CY

[schneider2037]

Git it. Thanks for the helpful link!

[schneider2037]

Gotcha. It also looks like we're fully reloading transaction facts - which is crazy. Why reload transactions that haven't changed since 2012 every night!? Looking to switch the underlying models to incremental. Methinks here lies the major part of the overnight performance hits.

Thanks!