**Architectural Logic**: Normalization ensures integrity; SCD-1/3 handles dimension updates. **Normalization**: 1NF–3NF; reduces redundancy; consistency; efficient updates. **SCD-1**: `UPDATE dim SET col = new_val WHERE key = X`. **SCD-3**: `UPDATE dim SET prev_col = col, col = new_val, updated = NOW() WHERE key = X`. **MERGE**: `MERGE dim USING staging ON dim.key = staging.key WHEN MATCHED THEN UPDATE SET ...`....
The complete answer continues with detailed implementation patterns, architectural trade-offs, and production-grade considerations. It covers performance optimization strategies, common pitfalls to avoid, and real-world examples from companies like Meesho. The answer also includes follow-up discussion points that interviewers commonly explore.
Continue Reading the Full Answer
Unlock the complete expert answer with code examples, trade-offs, and pro tips - plus 1,863+ more.
Or upgrade to Platform Pro - $39
Engineers who used these answers got offers at
AmazonDatabricksSnowflakeGoogleMeta
According to DataEngPrep.tech, this is one of the most frequently asked SQL interview questions, reported at 1 company. DataEngPrep.tech maintains a curated database of 1,863+ real data engineering interview questions across 7 categories, verified by industry professionals.