Migrate relational-style data from NoSQL to Amazon Aurora DSQL

This article demonstrates migrating relational-style data from MongoDB NoSQL to Amazon Aurora DSQL using Kiro CLI and Apache Spark.

• Aurora DSQL offers serverless distributed SQL with ACID compliance and multi-Region capabilities
• Solution uses Kiro CLI generative AI to optimize schema design and generate DDL statements
• Apache Spark on Amazon EMR performs ETL transformation from denormalized to normalized relational schema
• MongoDB Spark Connector reads source data; PostgreSQL JDBC writes to Aurora DSQL
• Process includes entity identification, table creation, ETL script generation, execution, and validation
• Aurora DSQL limitations: no foreign keys, triggers, stored procedures, sequences, or auto-increment IDs
• CloudFormation template deploys MongoDB, EMR cluster, Aurora DSQL, and supporting infrastructure
• Validation includes entity count checks and row-level data comparison between source and target

The migration demonstrates how organizations can transition from NoSQL flexibility to relational database robustness while maintaining scalability through Aurora DSQL's serverless architecture.