Showing results for 
Search instead for 
Did you mean: 

Lumira Designer Filter Bar parallel processing

0 Kudos


I'm using Lumira Designer 2.2 and I'm wondering if anyone out there has found a way to filter multiple data sources in a Lumira Designer application in parallel (taking advantage of the processing groups), after a selection has been made in a Filter Bar.

For example:

1) Say I have an application with 6 data sources, each of them assigned to a different processing group. When I run the application, it takes advantage of the processing groups, loading the data sources in parallel, so the load time is fast.

2) My application also has a Filter Bar component, which has the 6 data sources set as Target Data Sources.

3) Once I select a filter in the Filter Bar and select "Go", it applies the filter to all 6 data sources but appears to be executing that serially instead. I'm not sure how to confirm that this is in fact the behavior, but it does appear to lengthen the time to filter as I add each Target Data Source (1 source is fast, 2 is a bit slower, etc.).

This can cause significant slowness on our dashboard when a filter is changed. If anyone has dealt with anything similar, I'd be curious to know if you found a way to improve performance.


Accepted Solutions (0)

Answers (1)

Answers (1)

Active Contributor
0 Kudos

Hello Stacey,

the filter component used should not influence Parallel Processing. In the moment when you apply a filter, it should be stored - but parallel processing would become relevant ones the result set is retrieved - which should be in parallel.

I would recommend to run the app with PROFILING=X to see when the result set are retrieved and if this happens in parallel.

Best regards,


0 Kudos

Hi Reiner,

Thanks for your comment. I did actually use the Profiling feature to find out that the SETFILTER operation, which applies the filter, is running in sequence and SAP support confirmed that is the expected behavior. Interestingly though, the rendering of the datasources, after the SETFILTER operation is complete, does seem to happen in parallel. So parallel processing is still partly helping us in this scenario, but because the SETFILTER isn't included in that parallel processing, we are seeing that affect the refresh time.

The other factor here is that the dimension we are filtering on is a hierarchy and it does appear that it takes significantly longer to perform the SETFILTER on a hierarchy vs a standard dimension. So if it takes 1 second to set the filter on each datasource and we have 10+ datasources, it starts to add up.

The ideal solution would be to reduce the number of datasources to filter, but because it's very difficult to make a reusable datasource across different visualizations in Lumira Designer, this is not really an option for us.