SAA-C03 ยท Data ingestion, lakes and analytics
S3 data lakes and AWS Lake Formation
Building a data lake on S3, zones and partitioning, and Lake Formation for fine-grained permissions on catalog databases, tables and columns.
Key points
- Data lake: data in S3 in zones (raw, cleaned, curated), partitioned Hive-style (year=2026/month=09/).
- Table definitions: AWS Glue Data Catalog.
- Lake Formation: permissions on top of IAM for databases, tables, columns; data filters for rows and cells.
- Enforced in Athena, Redshift Spectrum, EMR, Glue; LF-Tags scale grants; cross-account sharing; audited in CloudTrail.
Source: https://docs.aws.amazon.com/lake-formation/latest/dg/what-is-lake-formation.html
Practice questions
What is the finest level at which AWS Lake Formation can restrict access to data lake tables?
- Only whole S3 buckets
- S3 prefixes, but not individual tables
- Tables, but not columns or rows
- Columns, rows and individual cells
Show answer
Answer: D. Columns, rows and individual cells
Lake Formation permissions reach down to columns, and data filters add row-level and cell-level control. Bucket-level or prefix-level control is what plain S3 policies give you, which is much coarser.
Which practices help build a well-organized data lake on Amazon S3?
- Keep raw, cleaned and curated data in separate zones
- Overwrite raw files after each transformation to save space
- Partition data by date with paths like year=2026/month=09/
- Store files of two different tables under the same prefix
- Keep the lake on EBS volumes attached to an analytics host
Show answer
Answer: A. Keep raw, cleaned and curated data in separate zones, C. Partition data by date with paths like year=2026/month=09/
Separate zones keep raw data untouched while cleaned and curated copies serve analytics, and Hive-style date partitions let engines skip data. Overwriting raw files removes the ability to reprocess, putting two tables' files under one prefix confuses crawlers and Athena, and EBS volumes are not the storage layer of an S3 data lake.
Once you adopt AWS Lake Formation, IAM policies no longer play any part in access to the data lake.
Show answer
Answer: False
False. Lake Formation's permissions model augments IAM rather than replacing it. Principals are still IAM users and roles, and hybrid access mode even lets you keep IAM-based S3 and Glue permissions for some workloads while moving others to Lake Formation.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
