Databricks Data Engineer Associate Question 247
Select 4A data engineer is working on a Databricks project and needs to decide between creating a managed table or an external table for a dataset stored in an external cloud storage location. Which of the following statements accurately describe the differences between managed and external tables in Databricks?
- A
Managed tables store both the data and metadata in the Databricks-managed storage location.
- B
External tables allow the data to remain in its original location, but the metadata is still managed by Databricks.
- C
Deleting a managed table will delete both the metadata and the underlying data from storage.
- D
Deleting an external table will delete the metadata but leave the underlying data in its original location.
- E
Managed tables offer better performance than external tables for all workloads.
Show answer and explanation
Correct answers: A, B, C, D
Explanation
Managed tables in Databricks are fully controlled by Databricks, with both data and metadata stored in its environment. External tables, on the other hand, reference data stored outside of Databricks while Databricks manages the metadata. Understanding the behavior of each type, particularly around data deletion, is crucial for data engineers to make informed decisions based on their project requirements.
- A. Correct.
Correct. In managed tables, both the data and metadata are fully managed and stored in Databricks' storage location.
- B. Correct.
Correct. External tables reference data in an external storage location, but Databricks manages the metadata for the table.
- C. Correct.
Correct. When you delete a managed table, Databricks deletes both the metadata and the underlying data from its managed storage.
- D. Correct.
Correct. Deleting an external table only removes the metadata from Databricks but leaves the underlying data intact in its original location.
- E. Incorrect.
Incorrect. The performance of managed and external tables depends on various factors, such as the underlying storage system and workload, but managed tables do not inherently guarantee better performance for all workloads.