4. Preview CSV impact without importing it

Goal: Identify affected consumers before importing a change.

Before you start: Original Targets and Consumers, Candidate.csv and nullable-handles.

The idea you will use

Current findings describe assets. Comparison predicts the impact of candidate keys within the selected scope.

Predict before running

Does comparing a CSV without T1 delete T1 from the asset?

Build and check

Check which consumers would lose their destination before importing a candidate table through your normal workflow.

1. Select Targets and Consumers again, click Use selected DataTables, and load Fixtures/nullable-handles.json.
2. In the CSV panel, choose Targets as the candidate destination table and select Fixtures/Candidate.csv.
3. Once both inputs are selected, wait for the comparison to finish. The candidate CSV removes T1.
4. Check the two current example errors and two additional ROWTETHER-ROW-REMOVED warnings for C1 and C2, which use T1.
5. Open Targets: T1 should still be there. The preview reads row keys and headers; it does not import the CSV or run import hooks.
6. Click Clear comparison to return to the findings from the current tables.

Do not read the warnings as if the row had already been removed. Review every consumer in scope before deciding whether to import the file.

If the comparison does not start, check that Targets is in the selected table list, is the candidate table and the CSV still has the Name header. Do not import Candidate.csv.

Check your reasoning

See the explanation

No. Comparison does not import. C1 and C2 would be affected; C4 and C5 remain current errors.

Your turn: Record current issues separately from predicted impact and verify that Targets still contains T1.

Ready to continue when: you can support a decision without confusing those two sets.