Set Analysis Aggregation?

[fkuyTuyt]

Hello, I'm confused about something and need some help, please.

I want to show difference in sales in 2019 vs. 2018, so in my measure expression I have this:

=sum({$<[Usage_data.Date.autoCalendar.Year]={$(vDate1)}>}Qty)-sum({$<[Usage_data.Date.autoCalendar.Year]={$(vDate2)}>}Qty)

Where vDate1 = 2019 and vDate2 = 2018. It works, but if I want to calculate difference in percentages like =sum((A-B)/B), so in this case it would be :

=sum(({$<[Usage_data.Date.autoCalendar.Year]={$(vDate1)}>}Qty)-sum({$<[Usage_data.Date.autoCalendar.Year]={$(vDate2)}>}Qty)/sum({$<[Usage_data.Date.autoCalendar.Year]={$(vDate2)}>}Qty))

Then the calculation returns blanks. What am I missing? Even if I try to do this:

=sum(({$<[Usage_data.Date.autoCalendar.Year]={$(vDate1)}>}Qty)-sum({$<[Usage_data.Date.autoCalendar.Year]={$(vDate2)}>}Qty)/2)

It returns blanks. Also, when I do:

=sum({$<[Usage_data.Date.autoCalendar.Year]={$(vDate1)}>}Qty)) <- notice second right parenthesis, which is an error, the expression editor doesn't indicate an error, but the result of this calculation is 0. Is it a bug? It should indicate an error. 

Second question:

Is there a more elegant way to calculate the above, e.g. difference in sales in 2019 and 2018? 

[Gysbert_Wassenaar]

Try

=sum({$<[Usage_data.Date.autoCalendar.Year]={$(vDate1)}>}Qty)/sum({$<[Usage_data.Date.autoCalendar.Year]={$(vDate2)}>}Qty)-1

[fkuyTuyt]

I had to switch vDate1 to vDate2 and vDate2 to vDate1, but it works as expected. Thanks a lot. Can you explain why your approach works and mine doesn't? 

[Gysbert_Wassenaar]

You're nesting sums which in this case is simply wrong. You don't want to calculate a sum of sums. You simply want the variance of two different sums.

If you ever want to do a sum of sums you also need to use the aggr function to specify the set of fields over which the outer sum should aggregate the values.