Resolve a finding
Decide whether a finding is a data problem, a missing destination or a rule configuration issue before changing your project.
Classify the finding
Start with the code, severity and coverage. A Completed run can still contain errors; a Partial or Cancelled run is not a safe basis for a project decision. A cancelled run discards its unfinished snapshot, so run it again.
ROWTETHER-TABLE-MISSINGmeans the declared destination table is absent or unavailable.ROWTETHER-ROW-MISSINGmeans the destination exists but the referenced row key is absent.ROWTETHER-REFERENCE-EMPTYmeans a required reference has no value.ROWTETHER-REFERENCE-INVALIDmeans the key cannot be represented by Unreal.ROWTETHER-ROW-REMOVEDmeans the candidate CSV removes a row used by a selected consumer.
Inspect the source
Select the finding and read its source table, row, property, destination and evidence. Click Open or double-click the finding to open the source DataTable and select the affected row. Do not assume that the displayed destination is the only consumer: RowTether reports the selected scope.
Ask two questions:
- Is the source value wrong for the intended relationship?
- Is the relationship declaration wrong or incomplete?
For a Name or String field, confirm the column profile and its destination path. For a row handle, inspect the handle’s actual destination and key. For a game-rule finding, inspect the rule conditions and whether the row is applicable.
Choose the fix
Correct the smallest responsible input:
- Fix the source row when its key or handle is stale.
- Restore or rename the destination row only when that is the intended data change.
- Update the profile when the field is a genuine
Name/Stringreference that was not declared. - Change a rule only when its requirement is wrong. Disabling a rule removes a check; it does not repair the data.
- For
ROWTETHER-ROW-REMOVED, inspect every named consumer before importing the candidate CSV. The preview is read-only and does not import the file.
Never edit a table while a review is collecting. A source change invalidates the run; complete the edit first and start a fresh review.
Re-run and confirm
Run Review tables again after the change. Confirm the run is Completed, read coverage, and verify that the finding disappeared for the intended reason. If it disappears because a rule was disabled or a table was removed from the selection, the data has not been proven healthy.
For a candidate CSV, clear or keep the comparison deliberately and verify both current relationship findings and removal warnings. Export the completed report only after checking the scope and privacy of the generated JSON, CSV and HTML files.