Skippr Docs

Appearance

MongoDB CDC to MotherDuck Final-State Guide ​

June 2026

MongoDB to MotherDuck CDC can stay compact without becoming vague about document identity, restarts, or deletes.

Short Answer ​

MongoDB CDC to MotherDuck works by reading change-stream document mutations and reconciling them into a DuckDB-style warehouse table. Skippr captures insert, update, and delete operations, uses fullDocument updateLookup for update after images, and applies final-state logic in MotherDuck so each document key resolves to one latest valid row.

What changes for MotherDuck is the operating footprint rather than the data contract. You still need resume tokens, a stable document key, order-token guards, and tombstones for deletes. The benefit is that the destination can stay lightweight while still enforcing transactional final-state rules.

Why Teams Struggle with This ​

Small destination stacks do not remove source complexity. MongoDB document updates and deletes still need an explicit warehouse interpretation.

  • MongoDB change streams require replica-set or sharded-cluster deployment.
  • Resume tokens are mandatory for safe restart behavior.
  • Document identity should stay tied to _id unless a different business key is explicitly defined.
  • Delete operations need tombstone protection so older replays cannot restore removed documents.

How Skippr Handles It ​

Skippr uses MongoDB change streams directly and stores the committed resume token after each batch, so the pipeline resumes where it left off instead of polling for whole-collection differences.

On MotherDuck, Skippr creates _skippr_order_token and tombstone tables automatically and applies documented final-state reconciliation with DuckDB semantics. That gives small teams an integration that is compact but still precise.

  • MongoDB insert, update, and delete events map directly to CDC mutations.
  • fullDocument updateLookup makes update events useful for final-state loading.
  • Resume tokens keep restart behavior durable and explicit.
  • MotherDuck reconciliation stays guarded by order tokens and tombstones.

What the First Useful Version Looks Like ​

Choose this setup when MongoDB is the operational store and the warehouse path should stay lighter than a larger lakehouse deployment.

The integration stays trustworthy when the team treats document identity and delete safety as first-class concerns, even in a compact stack.