[rmartinsen-ucd]

The import seeds every code as unclassified, so on test the classification table is never empty, and load-classifications.py refused to run. As a result, test shows about 2,400 unclassified codes that the CSV already classifies.

The loader now:

• classifies rows that are still unset (IncludeInReport NULL), using the CSV values
• inserts codes from the CSV that are missing
• never changes a row that already has a classification, so values set in the app still win
• lists the kept rows whose values disagree with the CSV, for review

Nothing else changed. It works on an empty table as before, and running it again is a no-op.

I also checked that the CSV matches local DataDb exactly: there are no differing values, and the only extra local rows are unclassified codes added by the seed.

Test plan

• Scratch DB with 3,000 unset rows, one conflicting app edit, and one code not in the CSV: 4,696 rows inserted and 2,771 classified, the conflict kept and listed, and the extra code left alone. Afterward the table matched local DataDb, apart from the unclassified codes only the local seed adds.
• Second run: 0 inserted, 0 classified.

🤖 Generated with Claude Code

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: b8ec4814-8618-4bd4-a416-0303731c957a

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

_{Comment @coderabbitai help to get the list of available commands.}