When you drill down on a particular attribute, the totals of the unique count can be greater than the summary row because many value of the attribute could have the same value of the field on which you are doing the unique count but each row of the drill down should not be greater than the summary.
Let me show through an example. Assume the data is as follows
Product Id Region
If we do a unique count of the Product ID, it will show 3. But if we drill down by region the answer would be like
Region Unique Count
In this case, the sum of the Unique Count is 5 even though the Unique Count before drilling down was 3. Let me know if this answers the question.
Thanks Siva for the clarification. So what i am seeing is a little different.
Im using a formula that looks like this :
unique count ( if ( change type = "new") then accountkey else null ) - max( if ( change type != "new") then 1 else 0)
Which gives me the count I want .
but when I drill down, i see extra records that do not have a "Change type" of "New" . Im expecting to only see "New".
The extra records show a "0" which means that they don't have a change type of "new". But they are still included in the drill down.
Not sure if the explanation helps