AWS has introduced a new feature in its AWS Transform service that allows users to customize the conversion of Microsoft SQL Server databases to Amazon Aurora PostgreSQL. The key innovation is the transformation plan, a set of rules that governs how the conversion process works. This plan lets users make critical decisions upfront, such as how data types like MONEY should map between the two systems, whether schema names should be preserved or altered, and how case sensitivity should be handled.
The transformation plan is generated automatically after an initial assessment of the database, but users can edit most of its settings before any conversion begins. This flexibility helps avoid costly rework later in the process. The plan covers execution waves, job phases, conversion rules, and target provisioning. For example, users can adjust the order in which databases are converted, skip unnecessary phases, or override default type mappings.
One detailed example in the blog post explains how SQL Server character types map to PostgreSQL, demonstrating how technical settings can impact application behavior. The post also emphasizes the importance of reviewing the plan early to catch potential issues before they become embedded in the converted schema. This feature builds on AWS Transform's existing workflow, which includes assessment, conversion, validation, and deployment.