**Architectural Logic**: Streams = CDC; Tasks = scheduled SQL; together enable near-real-time ELT. **Streams**: Capture INSERT/UPDATE/DELETE from table; query stream for changed rows. **Tasks**: Scheduled SQL; can be triggered by stream data. **Pattern**: `CREATE STREAM s ON TABLE t`; `CREATE TASK process SCHEDULE='1 minute' WHEN SYSTEM$STREAM_HAS_DATA('s') AS INSERT INTO target SELECT * FROM s`. **Use**: Incremental processing; CDC pipelines....
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 Cognizant. 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.