Dash filter superseding cube filter

I have a data cube that uses a Top/Bottom block to report the most-recent entry from a table of transactions. That is, it sorts by date, groups by ID, and then gives me the most recent transaction for each ID by only returning the Top 1.

On the dashboard, I want to filter the results by other columns. So I added a filter. Not important, but the filter allows the user to pick which category of transactions they see.

Well I just learned that the filter on the dash is superseding the one in the cube. That is, when I pick a category, it goes and gets all transactions that match, THEN gets the most recent. I can’t have that.

I need it to FIRST get the most recent (Top/Bottom in the cube) and THEN filter by category on the dash.

How can I set it to do these in order? (I cannot do warehousing, I thought of that already.)

Do you have the category field exposed as a parameter in the cube (SQL SELECT/Tabular SELECT transform) and the dashboard filter connected to that exposed parameter? Can you instead connect the dashboard filter to the visualization itself with category as a slicer?

I’m not 100% sure I follow, but the dashboard filter is just connected to the category as displayed in the table (as a column, not as a slicer)

Would I have to create a copy of the category column to make it a slicer and then filter off of that copied column?

Hi Ken,

I can see something similar, and although I can’t confirm as this is not my own area, I think this may be a known issue or related to one (92096). In the interim or at least for now, can you try configuring the Top/Bottom transform or another intermediate transform to add an output element parameter on your category, filtered to All? From I see using version 10, you don’t need to do anything else - this parameter seems to prevent a dashboard filter from filtering the initial data before that transform.

Great solution! That worked perfectly! I just edited the Top/Bottom cube and added that parameter, and instantly my dash worked as intended. Thank you so much!

1 Like