SQL and Persistent Data

A duplicate close can become a false return if it enters a price panel unnoticed. A relational schema can reject duplicate instrument-date keys before estimation. This lesson stores fixture prices in SQLite, queries a bounded sample and reshapes it into a pandas panel.

• create a prices table with a composite primary key and useful columns;
• insert rows with executemany and parameter binding;
• query a date-bounded panel and pivot it to wide form in pandas;
• explain what "system of record" means for a research or risk pipeline.

1. Schema for prices
2. Load and query a panel
3. Why not only files?
4. Integrity checks in SQL
5. Parameter binding remains mandatory
6. Repository module sketch