Relational Migrator
Parent: MongoDB Migration Patterns · Published reference · snapshot 2026-10-23 · skill mongodb-cohort-f1ebe93a2c8b
↓ Facts as markdownall context files
Source-anchored research for Relational Migrator. Run status: COMPLETED.
These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.
How it works
- The Relational Migrator jobs page describes migration jobs as non-idempotent by default and says reruns insert new documents and embedded arrays; it documents `migrator.engine.transform.idempotency: true` for idempotent inserts and warns of substantial performance impact on large jobs. The 1.16 release notes qualify embedded-array behavior: mappings use idempotent writes by default when the source table has a primary key. (high confidence) (verified as of 2026-10-23) [source]
Facts and statements
- Relational Migrator maps a relational schema into a target MongoDB document model and transfers data to MongoDB; teams can start from a recommended schema, a one-to-one table-to-collection layout, or an empty model, then customize mapping rules. (high confidence) [source]
- Mapping rules can denormalize related tables by embedding rows as documents or arrays; MongoDB's example embeds Customer and OrderStatus documents and OrderLine arrays under an Order collection. (high confidence) [source]
- Teams can create and model a project from a relational DDL file or sample schema without a live source connection, but a live relational database connection is required to run a migration job. (high confidence) [source]
- The current Migrator runs snapshot jobs only, with each job transferring a point-in-time dataset and stopping; the 1.15 changelog moved continuous sync/CDC, LLM-assisted query conversion, entity code generation, time-series migrations, and custom JDBC connections into Application Modernization Platform engagements. The latest release notes reviewed were for 1.18.1, released 2026-09-10. (high confidence) (verified as of 2026-09-10) [source]
- DDL import parses table, key, and index statements and ignores other statements, so it is not a full database-object conversion path. (low confidence) [source]
- In a 2025 TPC-H demonstration, migration vendor Adiom used Relational Migrator to create a denormalized model, exported the .relmig project, and passed that mapping to its separate SQL batch connector; the article presents Migrator as a modeling step in a wider transfer pipeline. (low confidence) [source]
- Relational Migrator maps relational schemas into MongoDB document models with an entity-relationship diagram builder. Users can start from a recommended schema, a one-to-one mapping, or a custom model, and target Atlas or self-managed MongoDB. (high confidence) (verified as of 2026-10-23) [source]
- As of 2026-10-23, MongoDB says Relational Migrator runs only one-time snapshot jobs. Repeating snapshots through the REST API on a schedule produces separate jobs, not continuous sync. AWS Prescriptive Guidance and Estuary's Oracle guide still describe continuous replication or CDC, so use current MongoDB documentation to resolve that scope mismatch. (high confidence) (verified as of 2026-10-23) [source]
- Mapping rules can keep relational tables in separate MongoDB collections or embed child-table rows in parent documents and arrays. Calculated fields can derive target fields from relational source columns. (high confidence) (verified as of 2026-10-23) [source]
- In a 2024 TPC-H study, researchers used Relational Migrator to move PostgreSQL data to MongoDB and create embedded-document and mixed collection layouts. They reported that query conversion was incomplete and manually checked and edited converted queries. (low confidence) [source]
Children
- No children recorded.