Check The Rows That Do Not Join Cleanly
A merged table can look tidy while combining separate IDs or dropping records. Compare the result against the exact key values and the rule for missing or conflicting matches.
This exercise uses fictional CSV snippets and a stated merge rule. It does not run an import or use real records.
Get The Pack
Give the AI only the test input. Keep the review key and worksheet separate, and retain the first answer before requesting corrections.
Download comparison worksheet (.txt)
Check Three Merge Boundaries
| Case | Draft error | What to preserve |
|---|---|---|
| MERGE-01 | IDs 007 and 7 were treated as the same number. | Match exact text. Keep both original rows and their distinct owners. |
| MERGE-02 | The merge chose one of two conflicting owners for K12. | Keep the base row with the owner unresolved and flag the conflict for review. |
| MERGE-03 | L3 was dropped because the owner table had no match. | Keep L3 and leave the owner blank. A missing match is not a reason to drop a base row. |
Check the keys and the affected fields, not just the total row count. An unchanged count can still hide an owner assigned to the wrong ID.
Compare A Later Review
Reuse the same records and merge rule. Compare which IDs, blanks and conflicts each answer preserves. Keep the original response before asking for a repair.
Return to the CSV merge pack or choose your next check.
For a real import, use an approved sample and inspect the resulting IDs and row count in the destination. The data extraction practice offers a related check for missing values before a merge.