View and analyze lineage
Screenshot note: Screenshots are illustrative. Controls or labels may remain untranslated in the demo release, and pages without a faithful English capture retain the source illustration. Control names and layouts can vary by release.
This guide covers graph operations used in Data Catalog, Lineage Map, and Application Scenarios: search, direction, level, depth, filters, exports, processing detail, code slicing, and comparisons.
Direction is upstream, downstream, or both relative to current objects. Level is object granularity. Depth is the number of relationship steps expanded. Changing one does not change the others. An undisplayed node may still exist.
Search lineage
Global search
Search schemas, entities, properties, or script files. Select multiple schemas, entities, or properties to view their full paths together. A script opens a detailed lineage view. Verify systems and namespaces before combining objects.
In-graph search
Find and highlight schemas, entities, properties, or tags already loaded on the canvas. In-graph search does not expand depth or load additional objects.
Choose direction
The current entity is the analysis reference. Upstream shows objects producing or affecting it. Downstream shows objects it may affect. Full lineage includes both directions. Recheck current entity after changing direction; multiple current entities also change export and impact scope.
Choose a level
Drill from platform or schema to entity and property level, or navigate upward. A schema organizes database objects; entities include tables, views, reports, and files; properties include fields; operators are computation or transformation steps.
| Level | Question |
|---|---|
| Schema | How does data move between systems or database areas? |
| Entity | Which tables, views, reports, or files depend on each other? |
| Property | Where does a field or metric originate and flow? |
| Operator | How do SQL steps compute or pass a field? |
Set depth
Depth controls successive groups of directly related objects. Start small and expand key paths gradually to keep large graphs readable.
Filter objects
Filter by entity type or tag, with single or multiple selection. The filter icon reports hidden items. For A > B > C, hiding B can display A > C. Clearing the filter restores the original view; it does not delete B or its relationships. Use filters for temporary tables, core tables, responsibility domains, and systems.
Canvas controls
| Control | Action |
|---|---|
| SQL | Show or hide SQL in script lineage |
| Zoom | Adjust object scale |
| Full screen | Expand workspace |
| Display | Control temporary objects, Chinese business names, and indirect relationships |
| Edit | Enter controlled manual editing |
Edit mode provides create, undo, redo, delete, and save. Delete applies only to objects and lines added in the current edit.
Export impact reports
Choose table or field granularity, upstream/downstream/full scope, object types, whether to exclude intermediate processing, and an available format. Reports can include objects hidden by display controls, so they may exceed the visible canvas.
Right-click an object and select Impact Analysis for its all-depth report. Multiple current entities produce separate reports. Without an explicit current entity, scope may include all graph objects. Verify the reference before exporting.
Processing detail and operator lineage
The second-level lineage view expands processing details beneath an object relationship graph; it is not an independent lineage dataset. Tables, fields, and operators support this view. Operator lineage connects computation steps to inputs and outputs. fx marks a calculation expression. Highlight an operator's upstream and downstream paths to inspect its relationships.
Check processing detail against full SQL, variables, and scheduling context.
Code slicing
Extract relevant processing fragments for a selected entity or property. Slicing changes the displayed analysis scope without editing source code.
To inspect upstream writes, select the entity, right-click Code Slicing, and choose input analysis. To inspect a field's downstream use, expand the entity, select the field or metric, open Code Slicing, and choose output analysis.
Slicing helps trace definitions across steps, but a focused fragment is not necessarily a complete executable script.
Compare code and lineage
In a space file's comparison view, inspect line-by-line code revisions with syntax and difference highlighting.
Lineage comparison highlights added, modified, and deleted nodes or relationships across two parsed results and lets you switch sides. Confirm comparable harvesting scope, namespace, and parsing rules before attributing differences to business changes.
The current English environment did not provide stable, distinct captures for operator detail, code slicing, and code comparison. Use the controls and verification points above, and confirm the exact layout in your deployed release.
Verify results
Check current entity, system, namespace, direction, adequate level and depth, hidden objects, consistency between fragments and full scripts, report scope, and comparable parsing conditions.
Related: exploration, scenarios, SQL lineage, administration.