I am learning to use Power Pivot and write DAX and I am working through a book I bought written by Rob Collie and Avichal Singh called "Power Pivot and Power BI".
Early on it explains what happens when using CALCULATE() in a measure. A key point of understanding is explained as follows:
If a filter argument acts on a column that IS already on the pivot, it will override the pivot context for that column
So in a simple table called "GSR" I have a series of invoices with an invoice date, a product and an amount. I have another column that converts all invoice dates to the last day of the month to gather them together. I have created a measure called "Total orders" that is counting the number of rows.
I have created a pivot of this data with:
- Products in the rows
- Month end in the columns (but I've actually chosen Month and Year from the automatic breakdown Power Pivot has done on my Month_End column)
- Orders in the values
This pivot renders correctly.
Now the issue:
I've created two slicers that are feeding off two disconnected tables; one containing month numbers, and one containing years. Based on the selection from each of these I have a measure that creates a month end using the EOMONTH(DATE(),0). This measure is called "Comparison_Month_End".
I then have another measure called "Compare_Orders" that contains the following:
CALCULATE([Total orders],FILTER(GSR,GSR[Month_End]=[Comparison_Month_End]))
The point of all of this - this is meant to get the orders from the GSR where the month and year match the slicer selection. I want this value to appear in the pivot for every month selected i.e. not filtered by the pivot.
It doesn't work, however. This seems to me to be counter to what the book says, which is that if the filter in the measure is applied to a column on the pivot (in this case Month_End), the pivot filter is overriden. So, for example, if I have 31-Mar-18 in a column on my pivot and the comparison month end is 31-Jan-18, I would expect to see the January orders within "Compare_Orders" sitting next to the March orders, the February orders, the December orders and so on, but it only appears next to January.
To me this is doing exactly the same as an example in the book where there is a pivot with Year in the rows, Total Sales in column 1 and then 2002 sales in column 2. Column 1 shows sales for each year, column 2 shows the same sales figure (i.e. 2002 sales) against every row, even where the row year is 2001, 2003 or 2004. The 2002 Sales measure is using CALCULATE() summing Total Sales, filtering on "Sales[Year] = 2002".
Could anyone please explain why what I expected to happen doesn't seem to be happening please?
Thank you.
