Data modeling patterns for Amazon Quick Sight multi-dataset relationships

This article presents data modeling patterns for Amazon QuickSight Multi-Dataset Relationships, enabling users to combine multiple datasets at query time using inner joins.

• Simple star schema: central fact table related to multiple dimension tables for common analytics
• Snowflake schema: normalized dimension chains for hierarchical analysis across multiple join hops
• Galaxy/constellation schema: multiple fact tables sharing conformed dimensions for cross-process analytics
• Role-playing dimensions: single dimension table referenced multiple times in different analytical roles
• Multi-fact with different grain: fact tables at different detail levels automatically aggregated before joining
• Independent refresh schedules: each dataset refreshes separately based on data volatility (hourly to monthly)
• Row-level security at runtime: RLS policies enforced independently per dataset during joins
• Workarounds for circular joins, recursive hierarchies, ragged hierarchies, and parallel hierarchies via denormalization
• Current limitations: inner join only, no outer joins, 12 dataset maximum, no self-relationships, no SPICE/Direct Query mixing

QuickSight Multi-Dataset Relationships enable dimensional modeling at query time, supporting star and snowflake schemas natively while providing workarounds for advanced patterns through data modeling techniques.