In JMP 11, when making a Summary table, using the statistic "N Categories" actually counts rows that are marked as excluded -- meaning, there seems to be no way way to prevent rows from contributing to the count of categories in "N Categories". Is this a bug?
For example, try this table:
|Column 1||Column 2|
Exclude any of the rows. Then, go to Tables -> Summary, and take "N Categories" on one or both of the columns. In the resulting Summary table, you will see that "N Rows" is 2, but the numbers of categories are 3.
A simple workaround would be to select the non-excluded rows and make a Subset table (Tables -> Subset) of selected rows, or right-click "Selected" and click "Data View", then make the Summary table from there. Another way is to make cascaded (recursive) summary tables by getting the category columns you want to be unique first, and then using regular "N Rows" when aggregating over those columns. (The latter scales to taking categories of multiple-column keys as well as including missing values, which "N Categories" doesn't.)
I just wanted to confirm that the fact that "N Categories" includes excluded rows is a bug?
Summary appears to ignore the row state. When I need to do calculations only on non-excluded rows I use Tabulate.