Scope
This is not a context-free performance ranking. Future experiments will pin data size, query shape, hardware, and versions before comparing:
- Write and data-organization models
- Partitioning, ordering, and data skipping
- Aggregation, joins, and parallel execution
- Ecosystem, operations, and failure recovery
Initial view
Choose from update patterns, interactive-query needs, and the team’s operational boundary. Concrete conclusions will follow reproducible experiments.