Which architecture best supports creating a relational data store on OneLake with SQL CREATE statements?

• A. Data Lake
• B. Data Mart
• C. Lakehouse
• D. Data Warehouse

Answer: (C)

Relational data stores that you manage with SQL CREATE statements on top of a data lake are best supported by a Lakehouse. Lakehouse combines the scalability and cost-effectiveness of a data lake with a managed, relational SQL surface. It lets you define schemas and create tables using familiar SQL DDL, while storing the actual data as lake files (like Parquet) with transactional guarantees and consistent governance. That means you can issue CREATE TABLE statements to define structured, queryable tables directly over your lake data, enabling BI and analytics workflows without moving to a separate warehouse.

A data lake alone focuses on raw data without robust schema management or transactions, so it doesn’t natively support reliable SQL-defined relational structures. A data warehouse is a traditional, externally managed relational store but isn’t inherently built on top of the lake’s storage model in OneLake. A data mart is a subset-focused approach and doesn’t represent the full architecture needed for flexible, broad relational design on the lake. Lakehouse uniquely provides the SQL-centric, relational capabilities on lake storage that you’re targeting.