Filter/Hide Columns/Rows in Visualization without Filtering the Underlying Data

User_R9RRP

Is there a way to "filter" or hide some elements of a visualization without filtering the underlying data?

This matters when running calculations are used.

Say I have column Date with values December 1, 2,3 with Revenue 10, 30, 20

I create a running total Revenue column, but I only want to keep the first and the last date for example:

Date_______To_Date_Revenue

12/01_______10

12/03_______60

If I simply use Date to filter out 12/02, the chart becomes

Date_______To_Date_Revenue

12/01_______10

12/03_______30

Dori S.-Oracle

It looks like what you want is to group 12/02 and 12/03 into a a single displayed label and sum up their values, which is then incorporated into the running sum. You would do this by creating a calculated date field with a case statement. The new date field's formula would be something like:

case Date when 12/02 then 12/03 else Date end

Then instead of including the original date field in the DV, you would use the new calculated Date field. The rest in your DV remains the same.

User_R9RRP

Thanks Dori, what I provided is a simple example to demonstrate my intention, but I hoped to have a more general method. In the example, it makes sense to use your case when trick, but other cases, it's not applicable.

Dori S.-Oracle

Even in your more complex example, I feel the solution is in grouping values that you want to aggregate, if your intent is to include the values corresponding to those rows (via a sum, for example), but not display the labels. You would group them into a specified value. This grouping could be done in the UI via a calculation or in the model or data source. If you want to provide a more complex example where you feel the grouping would not work, we could give it another shot.

Gianni Ceresa

Hi @User_R9RRP ,

Taking a step back, what product (name and version) are you using?

Are you working on a "classic" analysis or on DV workbook?

Where does your data come from? RPD/semantic model? A dataset?

All these influence the answer to your question, because there are ways to get there, but ... it depends :)