Hi,
Using the period over period option probably won’t do this if the previous period needs different filtering. If you’re able to move your time dimension from Rows to Slicers, there are ways you should be able to do this although this means the rows would need to refer to days using only attributes like Week of Year, Month of Year, Day of Month, etc. rather than referring to full dates in like in your current table.
With your time dimension in Slicers along with the attribute Hour of Day, beginning in version 8 adding a slicer comparison should allow you to filter these two slicers differently between two different measures:
- Measure 1: filter the time dimension to MTD and hour of day to 11am-5pm
- Measure 2: MTD offset by -12 months (using the Advanced option in the menu) and 24h
In version 7 you can only filter one slicer at a time in slicer comparison measures and I don’t think that would accomplish your goal, but there is another option. If you create two separate metric sets that each display this same data but filter the slicers differently like above, then you can use a formula visualization to combine them into a single result and add your % Change at the same time. Your formula visualization could be created using a formula like this one, for example:
($table1.Metric Set 1.Current$ / $table2.Metric Set 2.Previous$) - 1
(You can click on the table column to insert it into the formula instead of typing.)
I can think of another option if preferred using a dedicated data cube, by adding an expression to the select transform (if this is from a database) to get the hour of each raw date value. For SQL Server this could be DATEPART(hour, [Date]). Then a Filter transform can filter the data so that Date is Less Than Current Year or Hour is Between 11 and 17 (setting the transform’s Operator to Or).
Hope this works for you
