major-domo-database/.claude/sqlite-to-postgres/botched-sbaplayer-matching/CSV_REVIEW_INSTRUCTIONS.txt
Cal Corum 7130a1fd43 Postgres Migration
Migration documentation and scripts
2025-08-25 07:18:31 -05:00

50 lines
2.0 KiB
Plaintext

# CSV REVIEW INSTRUCTIONS
## File 1: unmatched_players_for_review_updated.csv
**Resolution Column Values:**
- `ACCEPT` - Use the suggested_sbaplayer_id (pre-filled for exact matches)
- `USE_SBA_123` - Use specific SbaPlayer ID 123 instead of suggestion
- `CREATE_NEW` - Create new SbaPlayer record for this player
- `SKIP` - Skip this player for now (won't be processed)
- `REVIEW` - Needs manual review (pre-filled for partial matches)
**Pre-filled Values:**
- Exact name matches are pre-filled as `ACCEPT`
- Partial matches are marked as `REVIEW`
- No suggestions are marked as `CREATE_NEW`
## File 2: high_risk_player_matches_updated.csv
**Resolution Column Values:**
- `MERGE_123_INTO_456` - Merge SbaPlayer 123 into SbaPlayer 456
- `DIFFERENT_PEOPLE` - These are actually different people, keep separate
- `USE_SBA_123` - For player matches, use this specific SbaPlayer ID
- `CREATE_NEW` - Create new SbaPlayer record
- `SKIP` - Skip this for now
**Pre-filled Logic for SbaPlayer Conflicts:**
- Different bbref_ids = `DIFFERENT_PEOPLE` (bbref_ids are globally unique)
- Same bbref_id = `MERGE` (definitely duplicates)
- One has bbref_id, one doesn't = `MERGE` suggestion (review needed)
- Neither has bbref_id + identical names = `MERGE` suggestion
- Player ambiguous matches pre-select the first suggested SbaPlayer
- Middle initial conflicts are marked as `DIFFERENT_PEOPLE`
## Important Notes:
- **bbref_ids are globally unique** - trust them completely
- If two SbaPlayers have different bbref_ids, they are different people
- If one has bbref_id and one doesn't, they might be the same person
## Next Steps:
1. Review and edit the resolution columns in both files
2. Save the files when done
3. Let Claude know you're ready to process the changes
## Common Patterns:
- bbref_id mismatches (like "HALP") should usually be `CREATE_NEW`
- Different bbref_ids = always different people
- Common names like "Carlos Martinez" need careful review
- Middle initials usually indicate different people