I think it would be nice if we could count the rows of a table/group even if the table is empty. Is there a good way to do this? (I understand why this is not working with GroupBy, although I could imagine a special case for these kind of situations in that node.)
Thanks, gabor

PS.: I have found a -fairly ugly- workaround: add a new row with some kind of (fix) content, then remove one from the concatenated table’s grouped values.

Hi Aborg,
I do not understand what you mean with counting the rows of an empty table?
Do you want to group the table by all columns and check how many dulicates exist per group?
Because as long as you have at least one column left use this one as aggregation column with the aggregation method count.
Bye,
Tobias

No I mean that the table has no rows. (Actually at some stages of iterations it might have rows.)
Here is my setup:
Input table:
±----±----+
|order| one |
±----±----+
±----±----+

I group by order, and want to count the ones. If there were no rows the result contains no rows. If there were at least one rows then the result contains at least one row. I would not mind if the order column would be a missing value, and the count of ones would be 0 in case of empty input table. (Another option might be to allow empty list of grouping columns to get only the number of rows/the first row/… .)
Hope this clearified a bit.
Thanks, gabor

Hi Gabor,
now I understand your problem. But the group by node will not support this behavior. This would lead to the generation of a new row that wasn’t there and would only make sense for the count option. For all other options it could not be distinguished anymore between an empty table and a table that contains only missing cells in the first row.
Bye,
Tobias