Reviewed by Aditya Kumar · Last reviewed 2026-03-24
**Pricing**: $5 per TB scanned. No charge for DDL, failed queries, or metadata. **Implication**: Cost is directly tied to bytes scanned. Full table scan on 10 TB = $50 per query. **Optimization**: Partitioning (by date, etc.) prunes; 1% of data = 1% cost. Columnar formats...
Pricing: $5 per TB scanned. No charge for DDL, failed queries, or metadata. Implication: Cost is directly tied to bytes scanned. Full table scan on 10 TB = $50 per query. Optimization: Partitioning (by date, etc.) prunes; 1% of data = 1% cost. Columnar formats (Parquet, ORC)—read only needed columns. Partition projection avoids metastore calls. Compress data. Scalability: At 100 TB scanned/month, $500. At 1 PB, $5,000. Optimization can cut 80%+. Best practice: Always partition; always use Parquet/ORC; avoid SELECT *; use partition projection for high-volume.
Pro-Move: 'We cut Athena cost 70% by partitioning by date + region and converting to Parquet—same queries, 30% of scan size.' Red Flag: Running SELECT * on unpartitioned JSON—will blow the Athena bill.
According to DataEngPrep.tech, this is one of the most frequently asked Cloud/Tools interview questions, reported at 1 company. DataEngPrep.tech maintains an editor-reviewed database of 1,863 data engineering interview questions across 7 categories.