Could you check if you have allot of unique value's in some of your data model fields. Document Properties > Tables # Distinct.
If you use allot of Set Analysis calculations you could consider too precalculate some of your expressions in your data model and aggregate it to an higher level.
Can you upload a screenshot from your table viewer?
Thanks for your response!
I definitely use a TON of Set Analysis, that was my first thought as well, and one I am discussing with our BI Data team.
I'm just wondering what you may be thinking of when you are asking about #Distinct? Are you thinking perhaps the data set is too weighty and needs more Distinct values? Just wondering.
I'll try to get a screen of the table view up.
It would help allot to reduce the number of distinct value's in your model if they ain't necessary to use.
For example recordid's or receiptnr's.
Set Analysis is a great tool when using flag's. But when used to often at 1 sheet it impacts the performance allot.
I would suggest to:
- move the calculations to the backend
- Aggregate data (from day to week for example)
- reduce the number of distinct value's
This would do the trick
There are a few things you can look into:
- RAM and CPU usage during those selections
- Does the user experience this issue all the time or at some random times
- What kinda compression are you using for saving the file (High,Medium,None)
- As suggested by Jelco where ever possible move calculations to backend
- Use where exists to bring in data thats truly needed
- Check subset ratio and information density in table viewer
- Try miminizing use of if() , count(distinct ), functions in charts
Since you mentioned you are experiencing this only on server might be to do with Server resources. Are there a lot of other apps which large data sets.
- You can also check QVS Statistics tab on QEMC or session/performance log files.
My hunch is that this is server related as well, as the qvw's themselves move like lightening when being used within the Qlikview Application, and only seem to experience these issues via use of the webserver/ajax/ie plugins.
Our datasets are all pretty large, so that could have affect. Unfortunately I am not the server admin, so I can't get readily at that info. This is all helpful though as it adds things I can review with the team!