Rolling 12 months sum

[phoenix]

Hi, I have below dataset with month_end date and Sales. I have to calculate Rolling 12 month sales for each month end.

I have tried

sum(aggr(rangesum(above( total sum({<month_end=>}sales),0,12)),month_end)) , but for 20140131, for past 12 it is picking up values for 20150831,0731.... in a round robin way.

for 20140131 I just want to see rolling 12 mon sum for sales = 100 . Are there any other ways to accomplish Rolling 12 month sum with these month end dates?

month_end	sales
20140131	100
20140228	101
20140331	102
20140430	103
20140531	104
20140630	105
20140731	106
20140831	107
20140930	108
20141031	109
20141130	110
20141231	111
20150131	112
20150228	113
20150331	114
20150430	115
20150531	116
20150630	117
20150731	118
20150831	119

[sinanozdemir]

Hi,

I might have an answer for you. I believe we may need to implement sorted Aggr() function so that when you sort the dimensions, the 12 month rolling sum won't change:

Let me know if this is what you are looking for.

Thanks