We have an instance of 2008 R2 which is running approx 6 cubes. The situation I am finding is as follows.
A user connects to the cube via MS Excel 2010 using the normal Data Sources method. Lets say they create a very simple pivot table...Accounts on the rows & Months on the columns with a measure giving us the total value for each Account. At the time of the user creating the report there were 11 months of data in available in the cube (Jan > Nov) so the pivot table reflects this in Excel. The user has applies a few filters (say 5) on the row to select some Accounts ad hoc. The user then saves this file to his desktop and leaves it until Decembers data has become available in the cube. December is upon us and we refresh the cube so that it now contains Decembers data as well. The user reopens the Excel file and right clicks on the pivot and refreshes the data with the expectation of seeing a full years data for the five account selected but this is not the case. What in actual fact happens is that all the data which is there before is lost and the pivot table is completely blank. Can anyone help with why this is happening and a potential solution?
I have an idea in my mind that Excel is storing the accounts being filtered using some unique identifier and when the cube is reprocessed for the new month new unique IDs are being assigned to the accounts and therefore a match is not happening. There must be a work around to this as otherwise every report which is generated is effectively a throw away effort. Cheers!