Usage of DistinctCount resp. DistinctCountover in a montly table

Question & Answer
1

My dataset is basically a time-series with IDs.
When I use distinctCount(ID) and create a table or graph, QuickSight seems to count distinct appearances of ID per time unit selected, not for the full population.
E.g. when an ID has one occurence in May and one in June, it should count it once in May, but not in June - however, as the sum of distinctcounts per months is bigger than the “total” distinctcounts, it seems to doublecount those.

I tired with distinctCountOver({ID},[{created_at}],PRE_FILTER) but this gives me the total number of occurences in each month.

Can you help?

2

Hi @ErikM,
Thank you for your question. ‘distinct_count()’ will count the unique values of a field grouped by any other field used in that visualization. It sounds like your visualization has a second dimension or a filter. You can add an added parameter in the function, separated by a comma, that can dictate the group by level. Please see the documentation below for reference.