MLA-C01 Data Preparation for Machine Learning Practice Question

A data scientist has a 40 GB CSV dataset in Amazon S3 that will be used to train a SageMaker model. The training script reads the data with pandas, and the scientist wants to reduce both storage cost and training-time I/O without changing the logical schema. Which data preparation action should be taken?

⚠ Common exam trap

The trap here is assuming that splitting or relocating files reduces cost, when only changing the storage format and compression actually shrinks the bytes stored and read.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

Convert the dataset to Parquet with Snappy compression and keep the same columns.

Parquet stores data by column and compresses each column efficiently, so a training script that needs a subset of columns reads far fewer bytes than it would from CSV. Snappy compression further reduces the S3 object size, lowering storage cost. Because the column names and types stay the same, the logical schema is unchanged and the pandas-based script can read the Parquet files with minimal modification.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
Move the dataset to Amazon EFS and mount it to the training container.
Why it's wrong here
Amazon EFS is a file system, not a compression or columnar storage format, so the data volume is unchanged. Mounting EFS also requires VPC configuration and adds per-GB throughput cost, which works against the cost-reduction goal. It does not address the row-oriented parsing overhead of CSV, so training I/O is not meaningfully improved.
✗
Convert the dataset to JSON Lines and enable S3 Transfer Acceleration.
Why it's wrong here
JSON Lines is still a text format, and its repeated key names typically make files larger than the original CSV, increasing storage cost rather than reducing it. S3 Transfer Acceleration speeds uploads over long distances but does not shrink the object or reduce the bytes read during training. This combination does not meet the cost and I/O objectives.
✗
Split the CSV into many smaller CSV files and keep the same column layout.
Why it's wrong here
Splitting a CSV into many smaller CSV files can improve parallelism, but it does not reduce storage footprint or eliminate the row-oriented parsing overhead. Each file still stores values as text with repeated column delimiters, so I/O volume and CPU cost for parsing remain high. This does not achieve the stated goal of reducing both storage cost and training-time I/O.
✓
Convert the dataset to Parquet with Snappy compression and keep the same columns.
Why this is correct
Parquet is a columnar format, so the training script reads only the columns it needs, and Snappy compression shrinks the on-disk footprint substantially compared with text CSV. The logical schema is preserved because column names and types remain the same. This directly reduces S3 storage cost and the bytes transferred during training, satisfying both requirements without changing the data model.

Quick reference

AWS S3 Storage Class Comparison

Storage Class	Min Duration	Retrieval	Use Case
S3 Standard	None	Immediate	Frequently accessed data
S3 Standard-IA	30 days	Immediate	Infrequent access, rapid retrieval
S3 One Zone-IA	30 days	Immediate	Non-critical infrequent data
S3 Intelligent-Tiering	None	Immediate–hours	Unknown or changing access patterns
S3 Glacier Instant	90 days	Milliseconds	Archive with instant retrieval
S3 Glacier Flexible	90 days	Minutes–hours	Archive, flexible retrieval
S3 Glacier Deep Archive	180 days	Hours	Long-term compliance archive

Go deeper

MLA-C01 Data Preparation for Machine Learning Practice Question

Why each option matters

Option-by-option breakdown

AWS S3 Storage Class Comparison

Related to this question