MLA-C01 Practice Question 73: DynamoDB Export Training Data
AWS Certified Machine Learning Engineer – Associate (MLA-C01) · Test 8 · question 3 of 10
An e-commerce company keeps customer order history in a 2 TB Amazon DynamoDB table that serves production traffic around the clock. Data scientists need a full, consistent snapshot of the table in Amazon S3 every week to build training data for a churn model. The database team insists that extracting the data must not consume the table's read capacity or affect its performance, and it wants as little custom code as possible.
How should the data be extracted?
Show answer and explanation
Answer:
DynamoDB export to S3 writes a table's data, as of any time within its point-in-time recovery (PITR) window, straight to an S3 bucket. Exports are asynchronous, do not consume read capacity units, and have no impact on table performance or availability. PITR must be enabled on the table. It is a managed feature that needs no scanning code, and each export is a consistent snapshot of the table at the chosen point in time (option A).
- B is wrong because a Scan, even a parallel one, reads through the table's read capacity. On a 2 TB production table that is exactly the load the database team ruled out. - C is wrong because Streams captures *changes* going forward. It never produces the existing 2 TB, and assembling a consistent weekly snapshot from a change log is custom work. - D is wrong because restoring a 2 TB table every week is slow and costly, and it still needs a custom scan script, when export produces the same snapshot directly.