DISTINCTCOUNT in Aggregation table?

[MarkEvansHW]

The composite data model using the "Manage aggregations" feature seems potentially very useful, but I'm really struggling with the lack of support for DISTINCTCOUNT.

It maybe the nature of the data I work with, but I never use COUNT because it is equal to COUNTROWS for me (source data doesn't have NULLs). I always, always, always use DISTINCTCOUNT.

There appears to be no way to implement an aggregation table with a distinct count summarization. If anyone has a solution or suggestion, let me know as this really makes the feature unusable for my purposes and is very frustrating as it appears it could be implemented. Almost every report, tab, visual I design has distinct counts.

[amitchandak]

DISTINCTCOUNT can not be aggregated it needs to be calculated from based data only. That is why not supported.

See if at a few places you can avoid that. Other you have to get it from the base table only. 

[MarkEvansHW]

Well, I can see that if additional summarization were needed, it wouldn't be able to use the DISTINCTCOUNT from the aggregation table but it would provide a high-performance result for common groupby combinations.

However, if the aggregation table has the exact value needed (same exact groupby values) then it could simply use the value that is directly stored in the aggregation table, just as it would for a SUM or a COUNTROWS. As with the others, the fallback would be a Direct Query.

In that sense, DISTINCTCOUNT could be aggregated and be helpful. As I said, I get that its value would be more limited than the others. If the visual needed something more aggregated, then it would have to use DQ to go back to the source, just as when a filter is added that isn't available in the aggregation table for SUM or COUNT.

Many of my reports are by day and an aggregation table with distinctcount would have the exact numbers I need.