A retail analytics company stores 40 TB of historical customer transaction logs in unlabeled .csv files across an Amazon S3 bucket. Several columns contain missing values, and the schema is not yet registered in the AWS Glue Data Catalog. An ML engineer must catalog the schema, clean and engineer features visually with minimal code, and store the prepared dataset for Amazon SageMaker training. 1. Run an AWS Glue crawler over the S3 bucket to infer schemas and populate the AWS Glue Data Catalog. 2. Run an AWS Glue DataBrew recipe job to impute missing values, apply feature encoding, and clean the dataset. 3. Write the transformed output dataset back to Amazon S3 in Apache Parquet format for SageMaker training. What is the correct chronological sequence of steps?
- A. 1 -> 2 -> 3
- B. 2 -> 1 -> 3
- C. 1 -> 3 -> 2
- D. 3 -> 1 -> 2
Reveal Verified Answer (A), Technical Explanation & Exam Hack
Correct Answer: Option A
First, the AWS Glue crawler inspects the raw CSV files in S3 to infer column schemas and register the table in the Glue Data Catalog (1). Second, AWS Glue DataBrew uses that cataloged dataset to run visual data cleaning and feature engineering recipes (2). Third, DataBrew outputs the cleaned dataset back to Amazon S3 in columnar Parquet format for downstream SageMaker training (3). Architectural Deep Dive: AWS Glue DataBrew is a visual, no-code data preparation service with 250+ pre-built transformations, automated data profiling, PII handling, and reusable recipes that data analysts and engineers can schedule across large S3, Redshift, and RDS datasets without writing Spark code. Why the distractors fail: Option B is incorrect because this option either misapplies the AWS service outside its intended architectural pattern or violates the primary constraint (operational simplicity, cost efficiency, or target accuracy) specified in the scenario. Option C is incorrect because this option either misapplies the AWS service outside its intended architectural pattern or violates the primary constraint (operational simplicity, cost efficiency, or target accuracy) specified in the scenario. Option D is incorrect because this option either misapplies the AWS service outside its intended architectural pattern or violates the primary constraint (operational simplicity, cost efficiency, or target accuracy) specified in the scenario.