A new multi-terabyte dataset needs an initial profile.
Inspect distributions and missing values cheaply before applying full checks to critical fields.
Which quality approach is most appropriate?
Use a representative, stratified sample for exploration and reserve full-data rules for fields that require complete coverage.
Detailed explanation
Transfer, time, and confidentiality risks increase unnecessarily.
Transfer, time, and confidentiality risks increase unnecessarily.
Deletion is not profiling and loses data.
Deletion is not profiling and loses data.
Sampling reduces exploratory cost while critical constraints retain complete validation.
Sampling reduces exploratory cost while critical constraints retain complete validation.
Ordering bias can hide anomalies and distribution changes.
Ordering bias can hide anomalies and distribution changes.
Check official documentation
Compare the following points with the official documentation.
AWS Glue Data Quality profiling、sampling、data statisticsの公式説明を確認する。
What to confirm
先頭行だけの確認と代表性を設計したサンプリングを区別できる。
Key points
- Sampling
- Representativeness
- Profiling
- Full-data check
Notes
- Environment: AWS公式DEA-C01試験ガイドと各サービスの公式ドキュメントの確認
- Confirm current AWS service behavior in the official DEA-C01 documentation.
Foundation review
Bias and error
Stratify by time, region, and category and record the sample method and uncertainty.
Operational split
Use sampling for exploration and complete checks for production quality gates.