Do not input private or sensitive data. View Qlik Privacy & Cookie Policy.
Skip to main content

Announcements
Join us in Toronto Sept 9th for Qlik's AI Reality Tour! Register Now
cancel
Showing results for 
Search instead for 
Did you mean: 
Not applicable

QlikView performance

hi,

my doubt is about performance the QlikView

what is the best way to get a sub-total during script load?
using "resident" or read the QVD file?

thank you

1 Solution

Accepted Solutions
Not applicable
Author

Hi Fabio,

reading from a qvd-file you can be done in two ways: optimized and "normal". In optimized mode it is similar to copying blockwise data (fast as lighning). But if you use a function like sum(), date() or what ever, QV has to read every row for evaluating and calculating. This is mostly faster then reading from excel or a database table, but as mentioned above, the optimize mode is the very fastest. So you have deside what kind of formatting etc. you can afford during creating your qvd to be able to use afterwards optimized mode from qvd to qvw. Another possibility is to load your data (independent from optimized qvd) and afterwards read from the resident table which is then in RAM. Your strategy depends on what functionality you want to achieve and from the data volume, because resident loading is also very fast, but you need the according memory at one time.

Regards, Roland

View solution in original post

4 Replies
Not applicable
Author

Resident It will give you a better preformance because the table is already in memory.

Not applicable
Author

Hi Fabio,

reading from a qvd-file you can be done in two ways: optimized and "normal". In optimized mode it is similar to copying blockwise data (fast as lighning). But if you use a function like sum(), date() or what ever, QV has to read every row for evaluating and calculating. This is mostly faster then reading from excel or a database table, but as mentioned above, the optimize mode is the very fastest. So you have deside what kind of formatting etc. you can afford during creating your qvd to be able to use afterwards optimized mode from qvd to qvw. Another possibility is to load your data (independent from optimized qvd) and afterwards read from the resident table which is then in RAM. Your strategy depends on what functionality you want to achieve and from the data volume, because resident loading is also very fast, but you need the according memory at one time.

Regards, Roland

Not applicable
Author

To summarize:

If you have enough memory: Use resident ( faster)

Else, read on the file and aggregate at the same time.

Rgds,

Sébastien

Not applicable
Author

Thank you everybody for all answers!