For near-real-time reporting with minimal development effort on a lakehouse with heavy historical data and frequent incremental updates, which storage mode is most suitable?

• A. Direct Lake
• B. DirectQuery
• C. Import
• D. Import and Direct Lake combined

Answer: (A)

Direct Lake mode is the best fit here because it lets you run queries directly against the data stored in the data lake, keeping data in place rather than duplicating it in a separate storage layer. This means:

• You see the latest updates without waiting for a data import or refresh, which is ideal for near-real-time reporting.

• It scales well with large historical datasets since you’re not moving a massive amount of data into an analytics engine; the data remains in the lake and is accessed as needed.

• Development effort is minimal because there’s no need to build and maintain import pipelines or incremental refresh logic; the lake’s files and metadata drive the query.

In contrast, importing data would involve copying a huge historical dataset into the analytics layer and then maintaining it with incremental refreshes to stay current, adding latency and maintenance. A Direct Query-like approach would query the source live, which can introduce latency and depend on the source’s performance and availability. Combining Import with Direct Lake adds complexity without offering a clear advantage for the given scenario.