If there are a lot of unique values in a single attribute it seems that the filter does not show option to show/hide

(empty value)

is this expected? Or is there a way to enable that option? Designer and dashboards seem to behave the same.

Hi Vaclav, the “empty value” option in a filter value selection is displayed when there are records in the dataset that have no value for the attribute being filtered. This could happen due to missing data and data integration issues where records may have missing or null values in the attribute in question.

This happens on VARCHAR column that has both null and not null values in the data. So it does not seem to be related to that.

Is the NULL value in the dimension dataset (where attribute is declared possibly as primary key)? Such dataset is used to obtain list of values for filter drop down and other datasets where attribute can be declared as foreign key referencing the dimension dataset are not taken into account. References in fact datasets are supposed to be subset of what is in the dimension dataset.

This specifically is just a column in a dataset (database table containing both facts and text attributes) that is of type VARCHAR and is not the primary key, nor reference to another dataset. The VARCHAR attribute has both

null

and text values. When I drag just a filter into the analytical designer (without any metric used).

May I ask which data source type you use? It may be the case that we read ordered data and null is returned as last. The API probably returns only 10000 first values matching the search condition (for performance reasons)

PostgreSQL

Thanks for reporting this. I can reproduce the problem. We use NULLS FIRST in some SQL dialects for ordering but it seems not for PostgreSQL for this use case. We will fix it.

Any idea about ESTIMATED timeframe when this could be fix?

The fix will be probably released on April 4th.