Snapshots of data - Can datamart solve this issue?

[CarlsBerg999]

Hi,

What i want to do is take snapshots of my data. My data source is Odata. Currently PowerBI doesn't support snapshots and the issue is that with each refresh, the new data overwrites the old data. What i want to do is take daily snapshots of my data and build a report on top of it to compare what changed between days 1 and 2. 

I watched some videos of Datamart and it seems to be a data warehouse (among other things). Can datamarts solve the "snapshot problem" that has been present so far in Power BI? How would this be done in data marts? I just made a data mart out of some data just to see what it is like and quite honestly i don't understand what the difference between a Dataset and Datamart is. I understand that Datamart does bring a lot of features online, but other than that this just seems to be an extra layer to perform the things we used to perform in Power BI Desktop. 

Any ideas on storing snapshots to datamart?

Thanks for all links, tutorials, ideas and help!

[TomMartens · Accepted solution]

Hey @CarlsBerg999 ,

indeed datamarts come with some very promising features, to get started take a look here: Dataflows with benefits – BI Polar (ssbipolar.com)

Snapshotting data can become quite challenging, no matter if you are using PowerQuery/Dataflows or datamarts. This is mainly because of the fact that whenever a query retrieves data existing data will be overwritten. This is how it works.
This basic principle can be overcome when you use incremental refresh, then the new chunks of data will be added to the already existing data, the new chunk(s) will be appended.
The problem that you have to solve is this: retrieve the same set of data (the same except changes, deletions, updates, or inserts and providing an extra field - snapshot date.

Maybe, this article by @GilbertQ might help you and provide some ideas on how to tackle your challenge: How to configure Incremental Refresh with Power Query in Power BI - Data Analytics | FourMoo and Pow...

Regards,

Tom