I find myself including the same data in almost all of my reports (Customer, Vendor, Material, Date, Plant, etc) and was wondering if there was a way to publish Each one of these datasets/datasources individually and then consume the published datasets when creating a new report.
For example If I create reports/model for sales, billing, and produciton orders then each one of these share common data (Material). When I publish these to the cloud and upload the data I'm essentially loading the same sets of data multiple times (50K materials for each report). I don't need 3 copies of my materials in the cloud (Same with customer, vendor, etc).
It would be nice If I could just model Materials and publish materials, then when creating a new report I would consume the published material dataset/datasource rather than pulling the materisl from sql server again thus duplciaitng the materil data that is being used.
Am I going about this the wrong way? When I'm creating my report I'm modeling it as a start schemea so I have my primary table query (Billign) and then I create queries for Customer, Plant, Date, and Material) in order to get additional details.
Rather than creating seperate material, customer, etc... queries should I be flattening that data into my primary fact table so that I only have 1 table and that table and that way I only have materials, customer, etc. that are required for my report? I have 56K material records but say for 2017 we may have only sold 6K different materials so from that stand point I don't need all 56K.
I'm commenting just to follow the responses to this. I've been curious about the same thing.
Often times, it makes sense to have a separate report for the sake of delineation of context (different business organizations, departments, etc). It's not to the complexity of needing a full-blown RLS setup, but definitely needs the separation.
The only way I've seen to do this is to create the subsequent reports IN the browser on PowerBI.com. This means that any new custom measures/columns have to be created in the initial PBIX that relates to that data source. I would love to have this ability in the Desktop version, but I'm not sure how it would work architecturally.
Ditto. There doesn't seem to be much info on this, however. If an Idea has been opened for this, I didn't see it. I know there is a work around manner of reusing Power Queries by sourcing from a text file, which is actually great for version control. I believe this discussion pertains to the same thing but for the entire PBI dataset as opposed to a single query.