How do you use transactions in PostgreSQL? What isolation levels does it offer and what anomalies does each prevent?
Show the explanation
A transaction groups statements so they succeed or fail together (atomicity). Use BEGIN … COMMIT, or ROLLBACK to undo; SAVEPOINT lets you roll back part of one.
begin;
update accounts set balance = balance - 500 where id = 1;
update accounts set balance = balance + 500 where id = 2;
commit; -- both changes become visible together, or neither doesPostgreSQL uses MVCC: readers see a snapshot and do not block writers, and writers do not block readers.
| Level | Prevents | Notes |
|---|---|---|
| Read Committed (default) | dirty reads | each statement sees data committed before it started |
| Repeatable Read | + non-repeatable reads, phantoms | the whole transaction sees one snapshot; may fail with a serialization error on conflicting updates |
| Serializable | + all anomalies (e.g. write skew) | behaves as if transactions ran one at a time; the application must retry on serialization failures |
Read Uncommitted is accepted but behaves as Read Committed — PostgreSQL never shows uncommitted data. For a "read then update" pattern under Read Committed, lock the row with SELECT … FOR UPDATE or do the change in one statement (set balance = balance - 500).
Original UniCircle question. Last reviewed 22 Oct 2026.