Skip to main content
Announcements
Have questions about Qlik Connect? Join us live on April 10th, at 11 AM ET: SIGN UP NOW
cancel
Showing results for 
Search instead for 
Did you mean: 
Not applicable

How to improve performance of section access documents

Hi,

For one of my clients on an extranet we have big documents (5GIG+) which are reduced by means of section access.

We would like to tune the document to its max. performance when using the ajax view.

Do you have any ideas / strategies to tune a big document after the document has been loaded into memory.

We already set the preload option and have install IIS on Windows 2012.

Kind Regards,

Dion

4 Replies
kaushiknsolanki
Partner Ambassador/MVP
Partner Ambassador/MVP

Hi,

     Section access is not going to reduce your data. It will only hide the data and will not give access to data to unauthorized users.

     Only way to improve performance is you look into data model for any improvements and delete unused fields.

Regards,

Kaushik Solanki

Please remember to hit the 'Like' button and for helpful answers and resolutions, click on the 'Accept As Solution' button. Cheers!
Not applicable
Author

Hi,

You need to modify your data model. Some ways to reduce the footprint of memory:

1) use numeric keys

2) reduce the number of the possible values of one field. If you have a field containing first and last names, cut it into 2 fields: one for the first name, one for the last name. If you have big data (for example from 1 to 1 million, you can get 2 fields: one for the thousands, one for the units => 2,125 will be stored as 2 in first field, 125 in second one)

The advantage of that technique is to "reuse" the pointers that QlikView creates for each value of the field. The main drawback is clearly to make the selections and expressions more difficult.

Moreover, the less different are the values in one field, the less memory each pointer will use.

Please read the excellent post of Henric: http://community.qlik.com/blogs/qlikviewdesignblog/2012/11/20/symbol-tables-and-bit-stuffed-pointers


Fabrice

qlikpahadi07
Specialist
Specialist

Section access  only hides the data size remains the same but  for such application size you can use QV Publisher which will reduce the data and create a different QVW apps of different size less than the original and mount  on QV Server on which you will get a good performance

Not applicable
Author

Hi Dion

I believe the document is reduced down upon opening by a user with Section Access (sorry to contradict you Kaushik).  If you look at Document Properties and 'Opening' tab, it does state 'Initial Data Reduction Based on Section Access'.  You can test this by giving access to a test user for one record and see the size of the file upon opening.

All of the above are good ways to reduce down your data.  Other things to look out for are large numbers of decimal places (particularly on %'s).  If you can round these to zero decimal places or 1 decimal place, you are reducing the number of unique fields to 100 (or 1000 for 1 dp).  Also, performance will be increased if you use bands for figures wherever possible, e.g. for %'s; 0-20%, 21-40+% etc. As this again reduces the number of unique fields.

Taking this unique field reduction further, remove individual ID's (document number, etc) if possible.

My final tip is to input Calculation Conditions against large objects.  This will mean they are not calculated and held in memory should you be on a different tab, however there is a trade-off between memory and processing for this.

I hope this helps.  Good luck with your model.

Steve