QuestionQ38

VerifiedData Preparation for Machine Learning (ML)

An ML engineer must process thousands of existing CSV objects as well as new CSV objects when they are uploaded. The CSV objects are kept in a central Amazon S3 bucket and contain the same number of columns. One column contains a transaction date. The ML engineer needs to query the data according to the transaction date.

Which solution meets these requirements with the LEAST operational overhead?

Explanation

Amazon Athena provides serverless SQL querying for data in Amazon S3. A CREATE TABLE AS SELECT (CTAS) operation creates a table from query results and can partition the output by the transaction-date column, allowing date-filtered queries while avoiding infrastructure and custom processing code. AWS Glue Spark jobs, S3 replication and Object Lambda, and Data Firehose with Lambda introduce substantially more operational components; Firehose is designed for real-time streaming delivery rather than processing existing S3 objects.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!