I am having a warning on a GroupBy node. It says “Group contains too many unique values” and it is skipped.
Thing is I have set up the “Maximum unique value per group” to 1.000.000 and I still get the same message. My data in that group are far less than that.
Any help or suggestions?
A shot in the dark. Your initial dataset contains indeed a lot of unique values. After filtering you have a dataset with less unique values you want to do a GroupBy on. In that case , see if the Domain Calculator node helps you out.
To give you a bit of context. I need to do a GroupBy using 10 columns. max unique values of these columns is 150. within these groups (where I do the unique count) I cannot see more than 500.000 unique values of the column x that I would like to calculate the unique Values. thus I set up the “Maximum unique value per group” to 1.000.000.
Still I get the warning, and to me this is a blocking data issue, that specific groups are skipped.
How can I overcome this issue?
How can the Domain calculator help me? to me it seems irrelevant.
Additional info: I have a loop for one of the grouped columns as I was getting a Java heap space issue.