Skip to main content
Announcements
Global Transformation Awards! Applications are now open. Submit Entry
cancel
Showing results for 
Search instead for 
Did you mean: 
Not applicable

QVD Size and performance


Hello Guys,

I am looking for a suggestion for qvd size and performance.I have two tables and I need to bring these data into qlikview.On joining these two tables causes cross joins and it is correct for my scenario and gives me correct set of data.

Only problem is the size of dataset i got.I got around 1crores of records.So it is better to bring whole data in one qvd or create two qvds(each per table) and join them on at run time.?

Thanks in advance.

5 Replies
Not applicable
Author

Without any details it is hard to say what is best for your load. In general it is best to load both QVDs optimized (no Transformation / where clause) as this is the fastest way to get the data into memory. After this you can Join / transform the tables in memory.

its_anandrjs

Better if you join this QVD and make single QVD from this QVD with cross join as you say and then use this QVD i believe it gives right result to you.

ashfaq_haseeb
Champion III
Champion III

It really depends on what you are trying to achieve. If you use join without link then it will create Cartesian product and your table will explode.

I would have kept them differently and use set analysis to do required analysis.

and try to join then with some link if possible.

Regards

ASHFAQ

tresesco
MVP
MVP

Creating one qvd could be a better option. That would improve performance because in the front-end on selection the data search would not require table hopping.

vardhancse
Specialist III
Specialist III

if you think there is huge data.

instead of splitting the QVD, go for splitting the QVW based on some condition some thing like region/quarter.

because if data is more in one single QVD, then the performance in the QVW will impact.

One more advantage is that if the same QVD was being used in another dashboard there will be no much impact for that dashboard.