Transactions and Isolation Levels

Prerequisites: PostgreSQL Fundamentals; Indexing and Query Optimization.

Isolation is the part of ACID that governs what one transaction is allowed to see of another transaction's in-progress, not-yet-committed work. Without any isolation at all, you'd be exposed to a menu of well-defined concurrency bugs: a dirty read (seeing another transaction's uncommitted changes, which might get rolled back), a non-repeatable read (re-reading the same row twice in one transaction and getting different values because someone else committed a change in between), and a phantom read (re-running the same range query twice and getting a different set of rows). Postgres's MVCC (multi-version concurrency control) design makes dirty reads impossible at any isolation level — every transaction sees a consistent snapshot, never raw uncommitted state — but the other two phenomena are very much possible depending on which isolation level you choose, and this project exists to make that concrete rather than theoretical.