GCP Cloud Digital Leader Practice Question

A data lake is designed to store very large volumes of data in its original, raw format, whether the data is structured, semi-structured, or unstructured. Keeping the data unchanged allows future teams to define schemas on read and apply different analytics or machine-learning techniques as requirements evolve. In contrast, a relational operational database is optimized for transactional processing of structured records, and a data warehouse expects cleansed, schema-on-write data for fast SQL analytics. A messaging service such as Pub/Sub moves data between systems in real time but is not meant for long-term storage or ad-hoc exploration. Therefore, a data lake best fits the scenario.