***Formatting was lost, should look like:

This approach can be more efficient, but for the sake of explaining i've left it a bit clearer....

• split the data into two branches, one for unique and one for duplicate (use a tester)
• Create a new field called unique for the unique branch and set the value as the count. Do the same for the duplicate branch
• Using an aggregator, set group by as the discipline and make sure its set to aggregate all fields

Thanks a bunch, I've just encountered this issue with the aggregator (I've isolated the issue to happen just after it runs).

I've noticed an issue with the aggregator settings: the "non-outlier" value is appearing in the outlier column. Can you help clarify this? I believed I set it up correctly, but I'm still new to aggregators.

Scaled version of workbench attached.

you've got two features with values in the 'outlier' field, so its taking the value of the first feature in and using that. As I can't see what the raw data is, i'm assuming the outlier field already exist in the input data.

Probably easiest to rename the outlier branch to something different, 'outlier_agg' for example, then after the aggregator rename it back to 'outlier'

thanks, this was very helpful.