Need Help With Identifying Error on Time Series

[kenyatta67]

Hi,

Good day! Would like your opinion on where I potentially have gone wrong in my model for time series to not work. My problem is that my current year and my delayed year are giving back the same results.

The code that I’m using is:

Policy Holders 2 =
CALCULATE(
DISTINCTCOUNT(Fact_MembersData[Serial Number]),
DATEADD(Fact_MembersData[Date].[Date],-1,YEAR))

For the delayed year

Policy Holders =
CALCULATE(
DISTINCTCOUNT(Fact_MembersData[Serial Number]))
For the original year

In terms of set up its just a simple match between the calendar table and the main data table. I have also nominated Date in the calendar table as my date table.

I’m really lost on why its not working.

[mspanic]

Hi @kenyatta67 ,

It seems )maybe I am wrong) that you don’t use you Calendar date dimension for that.

One of the main prerequisite for Time Series function to work is continous date dimension

Melissa made such great explanation in serial of posts about Time intelligence + provide code for continous Date dimension :

Extended Date Table (Power Query M function)

Hi @surpat, For Time Intelligence functions to work there are some requirements, you can find them in this recent post. Do I really need a full year in my Date table? - #2 by Melissa Alternatively I would suggest not creating a relationship between the Date- and Fact table but using TREATAS to create that relationship virtually since the granularities between the tables don’t match. As @Greg mentioned, if you have need of further assistance please create a new thread and provide all relevant…

If you have more dates /relationships to Calendar dimension you can use USERELATIONSHIP function.

Hope it helps.

Good luck.