Advance your Data & AI career with 50 days of live learning, dataviz contests, hands-on challenges, study groups & certifications and more!
Get registeredGet Fabric Certified for FREE during Fabric Data Days. Don't miss your chance! Learn more
Hi
I learned a lot from Matt Allington's article on golden datasets and how this is a best practice setup. I would like to ask about what best practice looks like in relation to this, but accomodating for a slight complication.
My company runs Sage 200 as its ERP. Within 6 months we will be making a decision as to whether to change to NetSuite or Dynamics 365.
I'm going to be building a lot of powerBI reports right now, but I don't necessarily want to have to rebuild them all when we change ERP. And for reports where we have long term trend analysis, it will be important that both Datasets are referenced.
So my question is this - what are the best practices here? What should I be thinking about right now in terms of how I setup my golden datasets so that our suite of reports can simply continue? Can anyone give me some advice and/or link me through to articles like matt's that detail this kind of scenario out?
Many thanks in advance
Stu
Thanks for that Ibendlin 🙂
I guess I was wondering if perhaps the answer to this was whether a data warehouse was the best practice. ie setup the desired table structure in a data warehouse and then feed data into it from the ERP1. Then when ERP2 comes along it's *just*(!) a matter of figuring out the queries to write to continue to feed data into that same structure. ie having the DW as an intermediate stage
So instead of :
ERP Data -> pbix dataset -> powerBI report
we'd have:
ERP1 Data -> DW -> pbix dataset -> PowerBI report
then:
ERP1 Data -> DW
+ append ERP2 Data -> DW
That feels like there wouldn't be a significant amount of rework (none?) to the [DW -> pbix dataset -> PowerBi] part when ERP2 comes along. Obviously there would be work required to figure out the [append ERP2 Data -> DW] part.
Keen to hear your/anyone's thoughts/feedback on that 🙂
A Datawarehouse is generally a good idea for decoupling production and reporting infrastructure. By itself it will not help you with the remapping - that is work you have to do in any case. It won't solve the problem for you.
And for reports where we have long term trend analysis, it will be important that both Datasets are referenced
Good luck with that. Whenever you change to a different hierarchy system and start to use different identifiers for customers, partners, products etc. you need to accept the fact that the hierarchy systems are different for a reason. That reason has nothing to do with rational thinking. It is entirely political and/or philosophical. It also means that your historical continuity WILL break. You may be able to continue to compare values on a very high level (country/region etc) but you have to expect it to break at the detailed level . You MUST prepare your management for that break and you have to expect them to not want to listen to you until the changeover actually happens, and then to yell at you for not telling them beforehand etc.
Advance your Data & AI career with 50 days of live learning, contests, hands-on challenges, study groups & certifications and more!
Check out the October 2025 Power BI update to learn about new features.
| User | Count |
|---|---|
| 5 | |
| 3 | |
| 1 | |
| 1 | |
| 1 |