I'm looking for a best practice as to how to handle changing measures when your source data changes.
We're in a (continual) fact-finding mode with our data. The extracts from our various ERP systems are complex, such that the names and logic of columns can change as we slice 'n' dice and come up with hopefully better extracts. Of course the complex measures from the prior downloads blow up when this happens, and they drop off the pivot tables and slicers as well. But, you can't change the measures prior to the new data since there won't be any columns for them to reference and the same thing happens! And if we name our modified columns with "old" names then we risk leaving incorrect measure logic in place.
So, just wondering if people have figured out a way around this to prevent having to reconstruct measures and pivot tables.
We're in a (continual) fact-finding mode with our data. The extracts from our various ERP systems are complex, such that the names and logic of columns can change as we slice 'n' dice and come up with hopefully better extracts. Of course the complex measures from the prior downloads blow up when this happens, and they drop off the pivot tables and slicers as well. But, you can't change the measures prior to the new data since there won't be any columns for them to reference and the same thing happens! And if we name our modified columns with "old" names then we risk leaving incorrect measure logic in place.
So, just wondering if people have figured out a way around this to prevent having to reconstruct measures and pivot tables.
