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: 
skyline01
Creator
Creator

Work-around for drillable dimension behavior

I am working in Qlik Sense 3.2.  I have a very simple app.  It has 1 internal table.  The fields are:

ReportMonthYear

TransactionAmount

Employee

Manager

Director

As you might guess, the Employee-Manager-Director relationship is a hierarchy, with the usual properties (i.e., 1 Director to many Managers, 1 Manager to many Employees).

I have 2 visualizations:

1)

line chart of Total TransactionAmount vs ReportMonthYear

2)

bar chart of Total TransactionAmount vs drillable dimension Director \ Manager \ Employee

My requirement is to make these 2 visualizations behave independently.  That is, selections in either visualization should not affect the other.  So, in my measure definitions, I have entered the following:

Sum({1} TransactionAmount)

However, I am noticing a particular situation in the data that violates this "visualization independence" requirement.  For a certain ReportMonthYear value, all rows in the table have the same Employee value (and, therefore, the same Manager and Director values as well).  That is, for the rows with this particular ReportMonthYear value, all rows have Employee = 'x', Manager = 'y', and Director = 'z'.  For this selection, Qlik automatically drills the dimension in the bar chart to the Employee level, which seems to contradict the purpose of the {1} set expression in its measure, and, of course, violates the "visualization independence" requirement.

Is there a way to make Qlik not automatically drill for this particular data scenario?  My proposed work-around is to split the internal table up into 2 copies, 1 for each visualization.  But, I prefer a different solution, as this could become very difficult to maintain if I have to create more visualizations.

6 Replies
Digvijay_Singh

See if this below has any link with your case, not sure though

Scatter Plot drilldown behavior

skyline01
Creator
Creator
Author

Thanks for the link!  Yes, that is the same thing as my situation.  It sounds like Qlik may be working on this "default engine behavior" in a future release.  As noted, one work-around is to ensure that there are at least 2 "values" by adding a null "value" for each level in the drillable dimension.  I think that is nearly impossible in practice for most real-world scenarios.  (For an app that frequently receives new data, am I to profile the new data to append null rows for each new scenario?)  So, perhaps my proposed work-around for my particular situation (replicate the data model / tables) really is the only solution (however clunky it may be).

Digvijay_Singh

mto

Hi Mike, FYI, One more person finding it ambiguous, Please see the link above having conversation with you on this.

krishna_2644
Specialist III
Specialist III

'requirement is to make these 2 visualizations behave independently' -  you can use alternate states .theres a workaround to create different states with help of p() function.


check this out -Using Alternate State in QLikSense (without Extensions) - hope that helps you

skyline01
Creator
Creator
Author

I don't necessarily need alternate states, as I have no need to compare states.  I simply need each visualization to behave independent of the others.  So, I don't need to use data islands + P function.  So, I think that my proposed solution (create independent copies of the only table in the data model, i.e., create a data island for each visualization) is the only one that is feasible at the moment.

krishna_2644
Specialist III
Specialist III

Sure. we can make that work too,by creating two different visualizations. all you need is to work on set expressions in both the objects.

please post sample data or qvf.

Thanks