Data type compatibility - MantrixFlow

For the live PostgreSQL-to-PostgreSQL workflow, keep destination columns explicit and conservative.

Recommended mappings

Source shape	Destination type
identifiers	`bigint`, `uuid`, or `text`
names and labels	`text`
amounts	`numeric`
timestamps	`timestamptz`
dates	`date`
booleans	`boolean`
nested JSON	`jsonb` or selected scalar columns

Casts

Use explicit casts in the destination-owned SQL transformation. The current workspace does not have a separate Normalisation step. Use SQL for structural conversions, null handling, JSON extraction, column renaming, or derived values. Save, validate, preview, and publish the revised transformation before the next run.

Destination validation

The destination table must contain every model output column you intend to deliver. Extra nullable destination columns are usually safe; missing required model columns are not.