Forum Discussion
Getting 'Refresh Only Unsynced Data' to Work Consistently in Power BI Desktop
Hey Power BI friends,
I’m Mark Pousee, and I’ve been tinkering with Power BI Desktop (in the Fabric context). I recently stumbled upon the “Refresh only unsynced data” feature sounds super handy for live datasets, but I can’t seem to get it to behave consistently. Most of the time, when I refresh, everything loads even the data I’ve already got. It almost feels like Live Query is forcing a full reload every time.
Here’s what I’ve attempted so far:
- Set up incremental refresh for my time-based tables, but refresh still pulls all historical data anew.
- Tried combinations of DirectQuery and Import modes—no luck.
- Enabled and disabled the setting multiple times, just in case it was stuck.
- Looked through Fabric community threads—no one seems to have a step-by-step sync fix.
Has anyone gotten this unsynced-only refresh to work reliably? Whether it’s a particular setup, version, or workaround, I’d love to hear what’s worked for you or if it’s just a glitch we’re all living with.
Thanks so much in advance!
Mark Pousee
Hi markpousee
In Power BI Desktop, the “Refresh only unsynced data” option doesn’t work as expected and often reloads the entire dataset. This happens because Desktop doesn’t maintain partitions the way the Service or Fabric does, so the feature is limited. It also doesn’t apply well with DirectQuery or mixed modes and is most useful only with Import tables. If you want a reliable partial refresh, the right approach is to set up incremental refresh using RangeStart and RangeEnd parameters and then publish to the Power BI Service. Once in the Service, partitions are respected and only new data is loaded, whereas in Desktop you’ll usually see a full refresh every time.
Hi markpousee,
Thank you for reaching out to the Microsoft fabric community forum. Also, thanks to rohit1991, for his inputs on this thread.
The key thing to know is that the “refresh only unsynced data” / incremental refresh policy does not fully kick in inside Desktop. You configure it there, but the actual partitioning that makes refresh incremental happens after you publish to the service and run the first full refresh. That first refresh creates the partitions, and then subsequent refreshes will only process the partitions that fall into your policy window. In Desktop, you will often see the full dataset reload even if incremental is enabled, which is expected.
If you are working with Direct Lake or Live Query, the refresh behaviour is a little different the model uses “reframe” logic to pick up file changes, and in some cases the service may pause automatic updates if there are errors. That can look like a full reload as well. Make sure the “Keep your Direct Lake data up to date” setting is enabled in the dataset and check query folding on your RangeStart/RangeEnd filters so the source can push those incremental filters down.
If you are not on Premium or PPU, a practical workaround is to keep historical data static and only append a smaller “recent data” table on refresh.
Refer these links:
Incremental refresh for semantic models in Power BI - Power BI | Microsoft Learn
Manage Direct Lake semantic models - Microsoft Fabric | Microsoft Learn
Hope this helps clarify things and let me know what you find after giving these steps a try happy to help you investigate this further.
Thank you for using the Microsoft Fabric Community Forum.
5 Replies
rohit1991Super User
Hi markpousee
In Power BI Desktop, the “Refresh only unsynced data” option doesn’t work as expected and often reloads the entire dataset. This happens because Desktop doesn’t maintain partitions the way the Service or Fabric does, so the feature is limited. It also doesn’t apply well with DirectQuery or mixed modes and is most useful only with Import tables. If you want a reliable partial refresh, the right approach is to set up incremental refresh using RangeStart and RangeEnd parameters and then publish to the Power BI Service. Once in the Service, partitions are respected and only new data is loaded, whereas in Desktop you’ll usually see a full refresh every time.
v-kpoloju-msftCommunity Support
Hi markpousee,
Thank you for reaching out to the Microsoft fabric community forum. Also, thanks to rohit1991, for his inputs on this thread.
The key thing to know is that the “refresh only unsynced data” / incremental refresh policy does not fully kick in inside Desktop. You configure it there, but the actual partitioning that makes refresh incremental happens after you publish to the service and run the first full refresh. That first refresh creates the partitions, and then subsequent refreshes will only process the partitions that fall into your policy window. In Desktop, you will often see the full dataset reload even if incremental is enabled, which is expected.
If you are working with Direct Lake or Live Query, the refresh behaviour is a little different the model uses “reframe” logic to pick up file changes, and in some cases the service may pause automatic updates if there are errors. That can look like a full reload as well. Make sure the “Keep your Direct Lake data up to date” setting is enabled in the dataset and check query folding on your RangeStart/RangeEnd filters so the source can push those incremental filters down.
If you are not on Premium or PPU, a practical workaround is to keep historical data static and only append a smaller “recent data” table on refresh.
Refer these links:
Incremental refresh for semantic models in Power BI - Power BI | Microsoft Learn
Manage Direct Lake semantic models - Microsoft Fabric | Microsoft Learn
Hope this helps clarify things and let me know what you find after giving these steps a try happy to help you investigate this further.
Thank you for using the Microsoft Fabric Community Forum.- v-kpoloju-msft
Hi markpousee,
Just checking in to see if the issue has been resolved on your end. If the earlier suggestions helped, that’s great to hear! And if you’re still facing challenges, feel free to share more details happy to assist further.Thank you.
- v-kpoloju-msft
Hi markpousee,
Hope you had a chance to try out the solution shared earlier. Let us know if anything needs further clarification or if there's an update from your side always here to help.Thank you.