Data transfer

The data-transfer component copies schema and data between any two data sources Dirigible knows about. Run from the IDE's Transfer view or driven programmatically.

When to use it

• Migrating an evaluation H2 deployment onto Postgres / MSSQL.
• Cloning a tenant from one cluster to another.
• Seeding a fresh data source from a reference one.

From the IDE

1. Open the Transfer view.
2. Pick source data source, target data source, source schema, target schema.
3. Optionally select a subset of tables.
4. Start. Progress, row counts, and any errors stream into the view.

The transfer is incremental on table-list - already-transferred tables are skipped on retry unless explicitly re-selected.

Schema vs data

The transfer first creates the target schema (DDL emission via the target's ISqlDialect) and then copies row data via JDBC. Constraints are added after data load to minimise insertion order issues.

Hooks

For programmatic transfers, plug in DataTransferCallbackHandler to observe per-table progress (onTableStart, onRowsTransferred, onTableFinish, onError). Useful for CI-driven migration jobs that need their own progress reporting.

Limits

• Cross-RDBMS type coercion is conservative - JDBC's getObject is used to keep portability. Custom column types (json, geometry) may need manual remediation post-transfer.
• For Mongo-backed data sources, transfer treats collections as tables; nested documents become opaque values. Use the Mongo SDK directly for high-fidelity migrations.

See also

• Transfer view
• Working with data
• Data anonymisation