Despite the fact that we have been able to do some crazy stuff with Aggr(), every time we wanted to explore the possibilities of inter-record or range functions we faced a wall, Aggr() array of values couldn’t be sorted by any means. You couldn’t control how the data was ‘stored’ in the aggr() virtual table (load order by default) making it hardly usable for a variety of cases such as cumulative calculations, inter-record and range calculations, etc.
An Aggr() array can contain virtually anything so you have to specify what data type will be sorted within the array of data (numeric, text, frequency or Load_Order) and the sort criteria, basically ASC or DESC. I strongly recommend you to check the Aggr() help page and get familiar with the new syntax.
(...) I needed to capture the percentage and number of the gains and losses in a KPI object(...) In order for the Aggr() function to work as I expected, I needed to make sure the MonthYear field was sorted properly in ascending order so I handled this in my script. Note that if the MonthYear field was not in ascending order, the Above() function may not always return the previous month thus returning the wrong results.
Now, to solve Jennell's dilemma we could simply use Aggr() avoiding any extra steps such as sorting the data in the script and making our calculation fully dynamic.