Author of a Towards Data Science piece revisits how analysts generate date tables for self‑service BI. For years he relied on DAX code inside Power BI models, copying a template that builds a calendar from the earliest to latest transaction date. The approach satisfies Microsoft requirement for full‑year coverage but forces maintenance in each model.

During a recent discussion the author discovered alternatives that shift the workload upstream. If a relational warehouse, Fabric Lake, or similar store already exists, the date dimension can be materialised there using either Power Query scripts or server‑side ETL, eliminating the need for DAX generation. Both methods produce identical columns—year, quarter, month, fiscal flags—and support locale‑specific formats via the FORMAT function.

The final option keeps the logic inside the model but leverages the newer user‑context‑aware calculated columns, allowing dynamic language rendering without duplicating columns. Choosing between DAX, Power Query, or warehouse‑level generation depends on data‑flow architecture: embed in the model for quick prototypes, push to the warehouse for enterprise consistency, or combine both for multilingual reporting. The article provides full code snippets for each path.