How to work with incremental refresh in power bi desktop and service?

[jaryszek]

Hello, 

I have sample data in power bi desktop. But want to use incremental refresh in service. 

I am using already bpip project with git support. 

How power bi workflow should work? Please correct me:

1. I am using StartRange and EndRAnge parameters to load only part of data into power bi desktop (small size)

2. Doing changes and saving it into bpip project

3. Publish with sample data but incremental refresh will take everything ? --> i do not have clarity how this is working. 
It will replace current report and once again load data into it? Or will be not looking at power bi desktop changes. 

Best,
Jacek

[jaryszek]

thank you!!

[jaineshp · Accepted solution]

Hey @jaryszek,

No, publishing won't reload all your data again. The service is smarter than that - it only updates the model structure and processes new increments based on your policy.

Main Points:

Desktop vs Service:

• Desktop: Use small date range (StartRange/EndRange) for development only
• Service: Follows your incremental refresh policy, not desktop parameters

Publishing Workflow:

1. First publish → Service loads full historical data per your policy
2. Later publishes → Only model changes update, existing data partitions remain
3. Regular refreshes → Only new increments load

What Actually Happens:

• Your desktop sample data gets replaced by full dataset on first publish
• Historical partitions survive when you publish model updates
• Service ignores your desktop date range and uses the refresh policy instead

Best Practice:

• Keep desktop data small for fast development
• Set proper incremental policy (e.g., store 2 years, refresh last 10 days)
• Publish once, then only new data gets processed incrementally

Bottom Line: Publishing won't reload everything - it's smart enough to preserve historical data and only process new increments.

That's it! The confusion comes from thinking desktop parameters control service behavior, but they don't.

Fixed? ✓ Mark it • Share it • Help others!

Best Regards,
Jainesh Poojara | Power BI Developer