Manage lineage spaces

A lineage space organizes resources, tasks, namespaces, and parsed revisions. It is neither a database nor a metadata catalog. A space groups work; a namespace identifies objects. Verify mappings even after files are added.

Resource types

Area	Content	Sources
Lineage Management	SQL, ETL, Excel files and results	Scripts, ETL, custom relationships
Database Resources	Harvested objects and models	Databases, DDL
Report Resources	Reports, datasets, preparation SQL	Report harvesting
OpenLineage Tasks	Run events and job relationships	Kafka/OpenLineage

Tasks can be created manually or synchronized from harvesting. Update harvested database and report objects at their source task. The Lineage Management resource area is distinct from the similarly named DAM metadata license.

Plan spaces and folders

Define business scope, source systems, environment, owner, and engine. Group spaces by domain or environment and folders by source and purpose. Keep development, test, and production scripts distinguishable. Configure roles, space permissions, and task permissions before shared write access.

Create lineage tasks

1. Open Lineage Space > Lineage Management > New Task.
2. Select folder, engine, and import method.
3. Configure full or incremental import, source folder, and metadata constraints.
4. Bind namespace, default schema, DBLink, and preprocessing.
5. Save and test a small sample before expanding.

Full import processes all files in scope. Incremental import compares content, size, path, and name. Forced binding requires parsed objects to match harvested metadata; an unmatched table or field can fail the task. Check coverage first.

Database and report resources

After database or DDL harvesting, check database, schema, table, and field counts and namespaces, then run parsing and inspect its report. Report resources contain reports, datasets, and preparation SQL. Resolve failures through connections, namespace mappings, and engine settings rather than direct metadata editing or deletion.

OpenLineage tasks

Verify event source, Job identifiers, event order, RunEvent times and states, parent-child relationships, and generated column lineage. Missing events should not be represented as observed runtime facts through manual edits.

Namespaces and schemas

Configure database type, default schema, connection, and DBLink mappings before parsing. Inspect object identity before correcting unbound namespaces, incorrect schemas, or inconsistent connection names. See namespace configuration.

Preprocessing and reports

Preprocessing normalizes variables, repairs format, or extracts embedded SQL using configured JavaScript, regular expressions, or models. Retain original files and script revisions. Reports distinguish success, partial success, syntax errors, missing metadata, and other errors. Inspect failed fragments and re-execute after correction; submission success is insufficient.

Check space quality

1. Review parsing success and partial counts.
2. Inspect coverage and unbound objects.
3. Investigate breaks, cycles, and isolated nodes.
4. Retain revision, time, and file traceability.
5. Have owners sample key table, field, and report paths.

Counts and coverage are clues, not proof of correct business lineage.

Revisions and corrections

Retain files and task records for automatic results. Use manual relationships only with evidence. Record objects, direction, reason, and owner before changes. Prefer reharvesting or reparsing to repair configuration, then remove incorrect manual relationships.

Troubleshooting

Symptom	Check
Successful task without a graph	Resource inclusion, namespace, partial report results
Tables without field paths	Field harvesting, SQL references, forced binding
Wrong same-name table	Namespace, schema, DBLink
Missing report lineage	Harvesting, connection, preparation SQL engine
Repeated incremental parsing	File path, name, size, content, and import mode

Related: resources, tasks, lineage views.