FIFO on millions of rows: how to reconcile bonuses, SQL, and an asymmetric N×M graph

▤ Summary

This article explains how to model FIFO allocation across bonus accruals and consumptions using SQL. It shows how to turn an N×M matching problem into cumulative ranges and solve it with interval overlap joins. The piece walks through CTE-based implementation patterns in Trino, Presto, and Spark SQL, and compares the approach with alternatives such as UNION ALL and any_value. It also highlights practical checks like reconciling accrued, spent, and burned amounts to validate the result.

▦ Linked Companies