# Data access and migration lab

Requires Python 3.10+, standard library only. Run in this folder. Fixture records are synthetic.

```bash
python run_checks.py --inventory
python sqlite_demo.py
python run_checks.py --template fixtures/pending_model.json submission.json --as-of 2026-09-12
python run_checks.py --check fixtures/pending_model.json submission.json
```

Between the last two commands, write your state, all typed issues, pointers, rationale, minimal repair and review trigger. STRUCTURE_PASS_SEMANTICS_UNREVIEWED proves shape only. Semantic correctness and explanation require review. Templates do not prefill the answer or overwrite existing files.

Read policy.json and the complete table/field map in resources. Work baseline and drop_add with the lesson, then the guided pending_model case. Independently analyze tracking_changed, unsupported_token, concurrency_conflict and transfer_provider. The example date is historical; record the actual date for a current claim and preserve expired states.

sqlite_demo.py actually executes SQL in memory only, not EF, SQL Server or a shared database migration. efcore_reference.cs contains teaching excerpts, not a buildable project. No real data or external access is needed. Preserve inputs and explain proposed repairs in your submission.