Momentum One Academy · glossary
Dataflow Gen2
Definition
Dataflow Gen2 is a low-code Microsoft Fabric item for connecting to data, applying repeatable transformations, and loading the results into one or more destinations. It uses the Power Query experience, so teams can clean, join, filter, aggregate, and reshape data without writing a full transformation program. 1
Where it sits in Fabric
Dataflow Gen2 belongs to the Data Factory workload and provides its visual transformation layer. It is not the same thing as a Data Pipeline: a pipeline coordinates activities, dependencies, schedules, and control flow, while a dataflow performs tabular data preparation. The two work together; a pipeline can trigger a dataflow activity, and the dataflow can write to Fabric Lakehouse tables or files, among other supported destinations. 1 2 3
When to use it
Use Dataflow Gen2 when a recurring ingestion or preparation step needs a maintainable, low-code transformation: for example, standardizing columns from several files, filtering records, or joining source data before analytics consumption. It is a practical fit when the output should land in a Lakehouse for notebooks, SQL queries, semantic models, or reports. Prefer a notebook, SQL, or other pipeline activity when the logic depends on custom code or procedural orchestration rather than Power Query transformations. 6
Practical example
A pipeline first copies public-holiday data into a Lakehouse. A Dataflow Gen2 then reads that table, filters Countryorregion to Belgium, and writes the result to a new Lakehouse table named BelgiumPublicHolidays. After publishing and refreshing the dataflow, the curated table is ready for reports or other downstream consumers. 5
Production caution
Choose the destination update semantics deliberately. For a new table, Dataflow Gen2's automatic settings use replace, managed mapping, and drop-and-recreate behavior. A refresh can therefore replace existing rows, and schema changes can recreate the table and remove relationships or measures added to it. Test this behavior with representative data, and use manual settings or a different update/schema strategy when the target is curated or must preserve those objects. 2
Module 3 connection
For Module 3, connect the three related ideas as one operating pattern: the Data Pipeline orchestrates the run, Dataflow Gen2 transforms the data into the Lakehouse, and the Monitoring Hub verifies status, duration, and failures. Monitor hub job runs explicitly support Dataflow Gen2, so operational checking belongs alongside authoring—not afterthoughts. 4
Ready to put this pattern into a governed delivery workflow? Continue with Momentum One Academy's Fabric Production Workshop.
References
Plan your workshop