Count Distinct Aggregation in Card Builder

I'm finding that a lot of customers would benefit from a Count Distinct Aggregation in the card builder. It is difficult to have drill down/filter information that is valid or useful in many situations when Count simply counts the number of rows.

 

Often in limiting results for one card to have accurate counts (at the data flow level), you eliminate information that is needed for drill down cards or undesirably change filter results. This necessitates lots of similar data flows and cards which are difficult for end users to manage or modify.

 

For instance, I have a customer who has a dataset where employees are counted on a monthly basis. Months on the X axis, employee counts on the y axis. Each employee can work for one or more properties, so the customer would like to drill down or filter to see what properties employees in each month worked for. This of course, would be only counting distict employeeIDs at the first card and then counting all employeeIDs once the card is filtered to a specific property. Since Count doesn't allow for distinct discrimination, this isn't feasible with only one master dataset.

7
7 votes

· Last Updated

Comments

  • Thank you for this idea. I am assigning this to our product manager @ckwright for review.

  • Thanks for this idea, we are looking into additional beast mode functions, we have this function on the road map.

  • I would like to second this feature, although it seems that it is possible today to create an arbitrary beast mode calculation using count distinct rather than a simple count.

    Specifically, I would like to see count distinct as an aggregation option when adding values/series to a card configuration, so that it is not necessary to create a beast mode calculation to achieve this goal.

This discussion has been closed.