What role do transaction isolation levels play in ensuring data consistency in a database?

Transaction isolation levels determine how much one transaction can see the changes made by other concurrent transactions, which directly affects data consistency.

Main isolation levels:

• Read Uncommitted: a transaction can see uncommitted changes of other transactions (dirty read).
• Read Committed: a transaction only sees committed changes, preventing dirty reads.
• Repeatable Read: guarantees that data read by a transaction will not change until it completes, preventing non-repeatable reads.
• Serializable: the strictest level, providing full isolation as if transactions were executed sequentially.

Choosing an isolation level is a balance between performance and data integrity. For example, in banking applications, Serializable is often used to prevent calculation errors.