You have a Fabric tenant that contains 30 CSV files in OneLake. The files are updated daily. You create a Microsoft Power BI semantic model named Model1 that uses the CSV files as a data source. You c

Question

Options

• AXMLA Endpoint is set to Read Only.
• BQuery folding is occurring.
• CThe data type of the column used to partition the data has changed.
• DOnly refresh complete days is selected.
• EQuery folding is NOT occurring.

• A
  3% (2)
• B
  6% (4)
• C
  2% (1)
• D
  73% (47)
• E
  16% (10)

Explanation

When 'Only refresh complete days' is selected in Power BI incremental refresh, the engine must process and validate full-day partitions for all 30 CSV files. CSV files in OneLake do not support query folding, meaning date/time filter predicates cannot be pushed back to the source. Instead, the entire dataset must be loaded into the engine's memory so filters can be applied in-memory. Combined with the 'complete days' requirement forcing evaluation of every historical partition to confirm completeness, this results in extreme memory consumption that exhausts F64 capacity resources. Option A (read-only XMLA) would block write operations but not cause resource exhaustion. Option E (no query folding) is a contributing factor for CSV sources but is not itself the configuration choice that triggers the failure - Option D is the specific setting that forces the expensive full-partition evaluation.

No community discussion yet for this question.