Schema discovery and custom types
Inspect source structure and turn reviewed fields or value libraries into reusable assets.
DataMaker can inspect connected databases and supported business-system metadata to help build templates and mappings. Start from Connections, select the source and inspect the available schema or service catalogue.
Review the structure
Check entity and field names, types, keys, required values and relationships. For databases, the available introspection depends on the enabled driver. MongoDB uses observed document fields or a collection validator and does not provide relational foreign keys.
For SAP, schema discovery reads OData metadata. Oracle Fusion uses its REST describe resources. A cached schema can become stale after a source upgrade; refresh or compare it before reusing a mapping against a changed system.
Refresh, compare, or find out what a change breaks
The graph DataMaker builds from a source's metadata is cached, so three separate operations exist and only one of them changes what you have stored.
| Operation | What it does |
|---|---|
POST /schema-graph/{id}/refresh | Fetch the endpoint's live metadata with its stored credentials, parse it, and replace the cache. |
POST /schema-graph/{id}/diff | Compare the cached graph against live metadata without touching the cache. |
POST /schema-graph/{id}/release-report | The same comparison, plus the templates and scenarios wired to that endpoint which the drift can affect. |
Compare before you refresh. A refresh overwrites the graph your reviewed templates and mappings were built against, and once it has, the evidence of what changed is gone.
The release report is the one to run after a vendor release - a quarterly Fusion update, an SAP transport. It answers the question a diff leaves open: not only what moved in the source, but which of your assets are pointed at the parts that moved. Review those before reusing them, and generate a small sample against the changed entity rather than assuming an unchanged field name means an unchanged field.
Choose how to use it
Database analysis can propose:
- Direct SQL fields, which read real source values through DB Response fields.
- Relational synthetic templates, which use structure and relationships to propose generated data.
Review the selected tables and dependencies. Choosing synthetic generation is different from querying source rows, and copying a schema does not copy its business rules automatically.
Reuse field configurations
A custom data type stores a reusable field configuration. A value library is a named weighted pool represented as a Custom field configuration and can be installed in a Hub pack.
When discovering values from a source, review sampling limits, frequencies and sensitive content before saving the list. A sample's frequencies need not represent the full table. See Distributions for the supported weighted shape.
Validate the proposal
Generate a small sample and check required fields, valid values, foreign keys and target formats. Resolve missing or ambiguous metadata explicitly. Use mapping dry runs before moving a larger dataset.