Microsoft Fabric broadens Delta Lake support for faster Power BI dashboards
Microsoft updated Fabric documentation to confirm Power BI Direct Lake models now read Delta tables at reader version 2. The change adds support for deletion vectors, column name mappings, and limited liquid clustering. Writers such as data warehouses and eventstreams can still write to the same Delta files in OneLake. Readers including the SQL endpoint and Direct Lake models gain access without copying data. Direct Lake remains read-only. Several Delta 3.x features and identity columns stay unsupported.
Until now, analysts facing slow imports or 200-step Power Query chains either accepted four-hour refreshes or duplicated data into separate models. The extra Delta features remove several technical blockers that previously forced those workarounds. The practical result is that a properly configured Direct Lake model can now deliver sub-second filter changes on larger tables without the usual import overhead or duplicated storage costs.
Analysis
Treat this as permission to stop wrestling inherited Power Query monsters on every new report. Open one modest table in Direct Lake mode this week and compare its load time against your current file.
Citation
This executive briefing was curated and analyzed by Collab365. To reference this analysis, please attribute: "This briefing is available on Collab365 Spaces (spaces.collab365.com)".