Configure namespaces, DBLink, and preprocessing

Namespaces distinguish identically named objects across systems and environments. Preprocessing transforms variables, connections, and irregular text into parseable input. Both affect correct object matching.

Plan namespaces

Open System Management > Lineage Management > Namespaces to create or edit unique names and attributes. Define technical source, environment, and lifecycle. SQL executed within a database usually uses its metadata namespace; cross-system ETL maps each source and target connection separately.

The default schema fills omitted schema names. Use the actual execution context rather than an arbitrary value to suppress missing-object errors.

Configure DBLink

Review harvested DBLinks or add one manually in DBLink mapping management.

Setting	Meaning
DBLink name	Connection keyword in code, such as remote in customer@remote
Default schema	Schema used when an object referenced through the link omits one
Creation namespace	Namespace of the database owning the DBLink
Target namespace	Namespace of the connected database

Test a cross-database sample after mapping.

Apply preprocessing

1. In Preprocessing Management, choose JavaScript, regular expression, or a large language model.
2. Define transformations for the database and code pattern. Configure model interfaces before model-based processing.
3. Select rules in the task's Preprocessing settings.
4. Check extracted variables and connections; export, maintain, and import them in bulk if needed.
5. Compare a small sample before and after processing. Preserve business meaning.
6. Parse and verify reports and lineage.

Models can extract SQL from Java or Python or suggest syntax corrections. Review output and retain originals; inferred rewrites are not evidence of actual system execution.

Diagnose broken paths

Check harvested metadata, all connection mappings, default schema, DBLink direction, variable replacement, and omitted statements. Reparse after corrections and verify results.