Schema evolution and DDL propagation

Keep structure moving with the data without making DDL invisible.

SQL Flow emits versioned schema events with row changes, updates decode plans, and applies compatible target changes automatically.

Talk through my use case Review compatibility

End-to-end events

Detect, persist, replay, and apply one versioned contract.

Readers emit normalized schema events, spool frames preserve them, and sinks consume them before dependent batches. Caches invalidate at the same boundary.

  • ✓Add, alter, rename, truncate, and drop event model
  • ✓Schema version attached to row batches
  • ✓Ordered replay after restart or relay
  • ✓Per-engine identifier and type mapping

Policy before mutation

Automate safe changes and surface the rest.

Operators choose whether compatible changes apply automatically, require approval, or stop. Destructive changes show source DDL, target impact, and recovery action.

  • ✓Automatic compatible target DDL
  • ✓Approval for policy-controlled changes
  • ✓Schema event history and audit
  • ✓Coordinated multi-master evolution

The payoff

See what the capability changes.

Ordered

DDL + DML

Schema versions travel with dependent changes.

Automatic

Compatible evolution

Additive and mapped changes propagate.

Fail safe

Destructive changes

Risky DDL stops at an approval boundary.

Take the next step

Make schema change part of the replication contract.

Bring representative DDL, mapping rules, and target constraints. We will map automation and approval.

Talk through my use case