Refactor the database schema to better model the data relationships:
Schema changes:
- Removed cards.card_type (redundant, identical to card_number)
- Removed transactions.card_id (unnecessary indirection)
- Added transactions.card_number (stores card number for all transactions)
- Made cards.customer_id NOT NULL (every card must belong to a customer)
- Made transactions.customer_id nullable (NULL for anonymized transactions)
Import logic changes:
- Only create cards for known customers (transactions with customer_number)
- Store card_number for ALL transactions (including anonymized)
- Skip cards/customer creation for anonymized transactions
Additional changes:
- Add 'db reset' command to drop and recreate database
- Update migration file with new schema
This simplifies queries and better reflects the data model:
- Cards table: authoritative mapping of card_number -> customer_id
- Transactions table: stores all raw data including anonymized cards
- Customer relationship via JOIN on card_number for known customers
Jakob
7a172c6fdb