Why DATEADD results appear in a different month

[Sasha]

Hi,

I've calculated the following measure:

Sales_Previous_Q = CALCULATE(DIVIDE(SUM(Data[Sales]),SUM(Data[Units])),DATEADD(Dates[Date_Key],-3,MONTH))

The thing is that when I put months on a table a long with the measure, I get the results on the months for shich (-3) refers to.

For example: if my data refers to Dec-Feb, the measure will calculate for Sep-Nov. I would expect to see the results in Dec-Feb as the measure is calculated compring to them. But I see them in Sep-Nov. What am I missing?

[tex628]

DATEADD moves the actual daterange a specific interval. So if you have a matrix with a column dimension of month, you are essentially placing a month filter on that specific cell. When you apply DATEADD to the measure calculated in that cell it moves the range the specified amount. 

So with no dateadd:

Jan-19	Feb-19
Sales in jan	Sales in feb

With dateadd -2 months:

Jan-19	Feb-19
Sales in nov	Sales in dec

If you want to display previous quarter, you need to make sure that the date table contains only the current quarter before you apply the dateadd to the column.

My advice is to do something like:
ALL(datetable),
Year = YEAR(TODAY())
Quarter = QUARTER(TODAY()),
DATEADD(datetable[dates],-3, MONTH)

[Sasha]

Hi,

in my example I see the sales under thier respective month. So if  current month is Jan I see previous month (Dec) under Dec, whereas I'd expected it to be shown under Jan as you said.

[tex628]

That sounds like you the issue is related to which dimension you are using in your table. 
The month column that you are using as the column dimension, is it from your sales table or your date table?