The Problem: Storing Only the Current State

Most services store data one way. A table holds one row per entity. Each change overwrites the old value. This lesson shows what that design loses, and why common fixes do not fully recover it.

Updates Destroy Information

Consider a customer account table with a balance column. An update in place is a write that replaces the stored value with a new one. Every UPDATE accounts SET balance = 40 statement works this way.

Say an account starts at 0. A deposit raises the balance to 60. A withdrawal lowers it to 40. The database ran three statements, but it kept only the last result.