Harvest custom metamodel data

DAM and SQLink both map instances to model structures but use different sources, roots, and lineage preparation.

DAM: plugins and templates

Install plugins

1. An administrator opens System Management > Plugin Management.
2. Install matching source and reverse engineering plugins.
3. Verify enablement before configuring the source.

METAMODELDEMO supports template validation. Confirm the production plugin choice with the delivery team.

Define a source and task

1. Create a custom source in Metadata > Data Sources.
2. Associate the system and plugin, then save.
3. Create a task through standard harvesting. It primarily maps source to model; the configuration name normally identifies the schema.

Generate and complete the Excel template

1. Download the model-based template from source details.
2. Preserve prefilled root ID and Name. Name corresponds to the schema.
3. Enter child data and parent ID/Name in child sheets.
4. Preserve consistent identifiers through deeper levels.
5. In Metadata > Datasets, select the source and Import Metadata.

Correct missing parent IDs, duplicate names, required values, and hierarchy errors in logs. Test a few objects, tree navigation, and groups before bulk import.

SQLink: file harvesting and synchronization

1. Select File in SQLink harvesting and configure the metamodel.
2. Enter the lineage namespace. The root ID may be empty; child Name values can express hierarchy.
3. Execute and inspect status and object counts.
4. Filter the SQLink catalog by source and confirm synchronization to DAM Datasets.

Do not apply DAM's required root IDs to SQLink files or SQLink's empty-root rules to DAM Excel templates.

Display and incremental updates

Verify root, children, property groups, relationships, dataset filters, and asset registration in that order. Submit only changed objects for incremental updates and retain identifiers to prevent duplicates.

Add custom object lineage

1. Create an Excel lineage task in a SQLink space.
2. Download the relationship template and enter sources, targets, namespaces, and schemas. Many-to-many relationships are supported.
3. Turn off Chinese display in object lineage analysis to inspect namespace and schema.
4. Import, parse, and verify dependencies from metadata objects.
5. For an empty graph, check synchronization to the space's database resources; synchronize manually if needed.

Metamodel relationships describe business associations; lineage describes actual processing paths. Neither replaces the other.