Need Help?

[qv_testing]

HI Community,

I have Huge data in my application, even performence very slow..

How to handle Optimization ???

I Know Some thing How to handle

Memory utilized by the qlikview objects and the calculations time can be monitored from the

document properties.

Minimized chart objects will consume less memory comparing to the maximized one, hence

use of autominimize option will be a good practice in this case.

Removing synthetic keys.

Droping temporery tables.

Handling Expressions and Calculations. (Using Set Analysis).

If i am say these techniquies, Interview prospective they are not satisfy.

Anything else??

We have to set any properties???

Thanks In Advance...!!!!

[MK_QSL]

1) Use Rob Wunderlich's Doc Analyzer and remove unwanted fields from your application.

2) If timestamp is there in your application is not used in your expression, split them in time and date

3) Use Autonumber function when associating two tables by merging more than one field

4) Use numbers while creating flag rather than text values as numbers are optimizing less memory

5) Try to use Star Schema instead of Snow Flake

6) Load only required data. i.e. If user want to analyse data for last three years, no need to load data for last 5 or 10 years.

7) Wherever possible, do all calculations at script level instead of doing dynamically at front end.

😎 Use Set Analysis instead of If in front end.

9) Don't display all charts at once in dashboard. You can show/hide charts using buttons

10) Try to sort the data in the script itself and use sort order as Load Order which is optimized one.

11) If it is necessary to load all data, try to improve RAM which is an essential factor in QlikView

There are many more ways to improve performance!

Happy Qliking !!!

[Chanty4u]

Hi

• Performance tuning means simply reducing the loading time of the charts and Dashboard loading time.  Performance tuning in Qlikview is done at various stages, ie.,in front end and backend of the qlikview application.
• Backend:
• Do all the required calculations in the script itself.
• Load only required data, for example if you are using only last 5 years data filter it out when you are loading data.
• Load only required columns and tables so that Qlikview file size would be reduced.
• If you are loading same tables in multiple dashboards then for the instance load from database and store it in QVDs and in other Dashboards load from QVD.
• Frontend:
• Use Set Analysis as much as possible.
• Don't display all charts in a single sheet or provide buttons to show and hide the charts.
• Don't display huge data in Tables, restrict user to do some selections so that the data would be reduced and load the table faster.
• Limit the calculations in front end and move it to scripting.

Thank you

Suresh

[aveeeeeee7en]

Hi

POINTS:

1) Use Staging Process - First Load the Raw Tables and Save it in Qvd >> Than use these qvds and implement new Logics as per your requirement and save it other qvds >> Finally, use these qvds in creating Data Model.

2) Use Binary Load.

3) Use Triggers - Suppose you have huge data from 2000 year to till date 2015.

So, you can use Trigger to display current Year and Current Month Data.

4) Do not use Calculated dimensions in Charts.

5) Not to use Complex Expressions - Try using Set Analysis.

6) Understand Data before Applying Left Join or Right as they sometimes Multiply Data to Huge Volume.

7) Handle Nulls.

😎 Do the complex Calculations in the Script instead of Front end.

9) Use variables for Hide & Show of Charts and for Expressions.

Apart from these Points you Know what all to do.................!!

Regards

Av7eN

[#5]

Performance Tuning for Loading Data:

1. Fetch only Required Field.
2. Explicitly convert the field data type (Sometimes It stores Int as Dual So If you convert it to INT It will be storing as INT)
3. Use Int fields to join tables
4. To decrease load time, you remove compression but note It will take more disk space.
5. Try to save QVD files on disk, not on network path

Performance tuning for App:

1. If possible combine more tables into 1
2. do table joins on numeric fields
3. Archive old data which is not required (If you maintain 2 years of data then you should remove data older than 2 years while everyday load )

[Siva_Sankar]

Optimization can be done in several levels.

1. Scripting

2. Designing or UI Level

3. Server Level

Whatever you said, you have to categorize it and say. Get the points like what are the factors that will affect the performance. Though RAM, CPU Cores, no of users accessing dashboard, network capacity are the factors influencing the dashboard performance, sometimes you don't have privileges to change these in your organization.

So as a developer, you can concentrate on scripting, designing and server level design considerations. Have to follow the best practices while designing the dashboard.

In case of tuning the existing dashboard, check whether a date field is used with timestamp. And the timestamp is not necessary, you can trim it and avoid the no of distinct values which is one of the way to improve the performance and decrease  size of the dashboard. Find and go through the attached documents and present it in a way that you work.