This example shows how mortality data can be reshaped into long format for both study-year and policy-year analyses. A SQL script creates sample exposure-and-events records with policy IDs, dates, and covariates, then cross-joins them with dimension tables of study years (2009–2015) and policy years (1–7). Using `inner join` with year-based date filters, it generates observation rows for each active policy period, while a separate join uses `datediff('year', ...)` to assign policy-year exposures. The code relies on standard SQL `create table`, `insert`, and `select` statements to demonstrate the data preparation logic.