Forum Discussion

Felisia's avatar
Felisia
Frequent Visitor
10 months ago
Solved

trying incremental refresh, but keep failed to load data.

My data source comes from Business Central on premise. I use Odata to load the table i need. and for 2 largest table (GLE & GLBudget Entries) i try to use incremental refresh using parameter. i also ...
Need Help
  • v-sshirivolu's avatar
    v-sshirivolu
    9 months ago

    Hi Felisia ,
    Since View Native Query is greyed out from the source, the OData feed and Business Central connector aren’t supporting query folding, which forces Power BI to pull the entire 23 GB table and breaks incremental refresh. This is a common limitation with on-prem OData setups where authentication or transformations disrupt folding early. Using an Azure Virtual Machine (VM) is still a good interim solution for stability and performance while you wait for Fabric access. The VM can host the On-premises Data Gateway, providing a secure bridge for Power BI Service to connect directly to Business Central. Although the VM itself won’t restore query folding, that depends on the connector, it will help improve refresh reliability and allow proper incremental refresh once folding is supported. Configure the gateway on the VM, reconnect through it in Power BI Desktop, and apply your date filter early in Power Query to minimize load. If needed, check gateway logs or firewall settings for connectivity issues.

rohit1991's avatar
rohit1991
Super User
10 months ago

Hi Felisia 

 

  1. The error usually happens because the dataset is too large or the incremental refresh setup isn’t correct.
  2. First, check query folding, Power BI must be able to push date filters (RangeStart/RangeEnd) to the OData source.
  3. Apply date filters early in Power Query, before merging or expanding any other tables.
  4. Test with a smaller date range (like one month) to make sure incremental refresh is working properly.
  5. Turn off refresh for tables that are only used for lookup or references.
  6. If some tables are too heavy, filter them down or split them into smaller tables.
  7. For large datasets, try using a Dataflow or Fabric Lakehouse for better performance.
  8. After changes, refresh manually in Desktop first, then publish and schedule it again.
  9. This step-by-step approach usually fixes the “failed to load data” issue