PowerBI Distinct Count Returns Incorrect Value

[t-zhguan]

Looking to see if anyone can provide some insight into this issue. I noticed a discrepancy between the distinct count function in Power BI and Kusto. Distinct Count in Power BI seems to have a threshold where afterwards, the result becomes an approximation/estimate.

As an example, I'm using a Kusto Query as my data source:

Table_Name
| where Status == "Success" and Timestamp<=todatetime("7-6-2018") and Timestamp>=todatetime("6-7-2018") 
| project Column1, Column2, Timestamp
| top 12647 by Timestamp 

When I'm using the distinct count function to get the number of distinct Column1 (strings of numbers and letters), the result is 1024 (which is the correct value). When I increase the rowcount to 12648, the distinct count value drops to 1020.

For larger rowcounts, at times, the PowerBI result would be higher, at times lower.

EDIT: Should mention that I'm using DirectQuery instead of Import. Just tried on Import mode and this does not appear to be an issue.

[Anonymous]

@t-zhguan,

Please open a ticket in Power BI support.


Regards,
Lydia

[Anonymous]

Facing a similar issue. Were you able to get it resolved ?