Forum Discussion
Best design choice for multiple granularity
Hi wilson_smyth,
Based on my research, I suggest you to use the merge funciton in query editor:
Referenc:http://radacad.com/append-vs-merge-in-power-bi-and-power-query
Also you could add an index column in each table, then you could merge them by the index.
Regards,
Daniel He
Thansk for the reply.
Merging/appending is not the issue. obtaining the data is easy via merge/append or just modifying my source queries. Im asking about design practices, specifically:
- when there are many shared dimensions, is it better to have seperate fact tables at the desired granularity so that i dont have to deal with nulls when a row does not exist at a higher granularity
OR
should i have 1 single fact table, and wherever there are NULLS due to data not available at a lower granularity, populate it with the "unknown" value that i populate into the dimensions & deal with it in dax code.
I find myseflf regularly in the same sitiuation. Therefor the subject interests me greatly
What was your end conclusion on this subject?Where you able to find more info?