Dataflows Gen2 Migration Guide: Upgrading Power BI Dataflows to Microsoft Fabric
Migrate from Power BI Dataflows Gen1 to Dataflows Gen2 in Fabric for enhanced performance, Spark transformations, and OneLake integration.
Dataflows Gen2 in Microsoft Fabric represents a major evolution from Power BI Dataflows Gen1, adding Apache Spark transformations, OneLake storage, and enterprise scalability. This migration guide covers assessment, conversion strategies, testing approaches, and rollback procedures. Our Fabric migration services have migrated hundreds of dataflows for enterprises, ensuring zero downtime and improved performance. Move confidently to the next generation of cloud data transformation while preserving existing business logic and schedules.
Frequently Asked Questions
What are the main differences between Dataflows Gen1 and Gen2?
Dataflows Gen2 (Fabric) adds several capabilities beyond Gen1 (Power BI): (1) Storage location—Gen2 stores data in OneLake as Delta tables (queryable from lakehouses), Gen1 stores in internal Azure Data Lake, (2) Transformation engine—Gen2 supports both Power Query mashup engine AND Apache Spark notebooks, Gen1 only supports Power Query, (3) Scheduling—Gen2 integrates with Fabric pipelines and job scheduler, Gen1 uses Power BI refresh schedules, (4) Monitoring—Gen2 provides detailed Spark logs and metrics, Gen1 has limited diagnostics, and (5) Capacity—Gen2 runs on Fabric capacity (any F-SKU), Gen1 requires Power BI Premium. Both support Power Query M language and incremental refresh. Gen2 is the future direction—Gen1 will continue working but receive no new features. Migration is one-way—you cannot downgrade Gen2 back to Gen1.
Will my existing dataflows stop working if I do not migrate to Gen2?
No, Power BI Dataflows Gen1 will continue functioning indefinitely—Microsoft has not announced any deprecation timeline. However, all new Dataflow features are only available in Gen2 (Spark transformations, OneLake integration, advanced scheduling). For organizations staying on Power BI Premium (not migrating to Fabric), Gen1 dataflows remain fully supported. Migration is recommended when: (1) Moving to Microsoft Fabric capacity, (2) Needing Spark-based transformations for complex data engineering, (3) Wanting to query dataflow output in lakehouses/warehouses, or (4) Requiring advanced monitoring and pipeline orchestration. If current Gen1 dataflows meet your needs and you are staying on Power BI Premium, migration is optional. Most organizations migrate as part of broader Fabric adoption, not due to Gen1 limitations.
How long does a typical dataflow migration from Gen1 to Gen2 take?
Simple dataflows (5-10 tables, basic Power Query transformations, no complex dependencies) migrate in 1-2 hours per dataflow. Complex dataflows (50+ tables, computed entities, incremental refresh, linked entities) require 1-2 days each. Process: (1) Export Gen1 dataflow as Power Query template (30 minutes), (2) Create new Gen2 dataflow in Fabric workspace (15 minutes), (3) Import template and reconnect data sources (1-2 hours), (4) Reconfigure incremental refresh and parameters (30 minutes), (5) Test refresh and validate output (1-2 hours), (6) Update downstream Power BI reports to use Gen2 dataflow (30 minutes per report). Enterprises with dozens of dataflows typically migrate 2-3 per week, completing full migration in 2-3 months. Parallel migration possible for independent dataflows. Allow extra time for testing—always validate row counts and data quality match Gen1 before retiring old dataflow.