Concurrency control keeps simultaneous transactions from corrupting each other.

Why It Matters

Many users can read and write the same data at the same time. Without control, results become inconsistent.

Dirty Read

Transaction B reads uncommitted data from Transaction A.

A updates balance but has not committed.
B reads that new balance.
A rolls back.
B saw data that never really existed.

Non-Repeatable Read

A transaction reads the same row twice and gets different values because another transaction updated and committed between reads.

Phantom Read

A transaction repeats a query and sees new rows because another transaction inserted matching rows.

Lost Update

Two transactions update the same value, and one update overwrites the other.

Techniques

• Locks.
• MVCC.
• Isolation levels.
• Optimistic concurrency.
• Pessimistic concurrency.

Interview Notes

• Dirty read involves uncommitted data.
• Non-repeatable read involves changed existing rows.
• Phantom read involves new or removed matching rows.
• Isolation levels decide which anomalies are allowed.
• Higher isolation improves correctness but can reduce concurrency.