Question

A data engineer at a retail company needs to build a daily sales summary in a Fabric lakehouse. The source is a large transactional table (billions of rows) stored as Delta in the bronze layer. The requirement is to produce a gold table that contains total sales amount and transaction count per store per product category per day, and the transformation must run efficiently over the full dataset using distributed compute. Which approach best meets these requirements?

Accepted Answer

Use a PySpark notebook to read the Delta table, apply groupBy on store, category, and date with sum and count aggregations, then write the result to the gold Delta table.

Answer

Use a Dataflow Gen2 with a Group By transformation on the bronze data and write the output to the gold lakehouse table.

Answer

Create a OneLake shortcut from the gold lakehouse to the bronze table and query it directly from Power BI to compute the aggregates.

Answer

Use the SQL analytics endpoint of the lakehouse to run a T-SQL view that aggregates the bronze table each time the report is opened.